پروف آف کانسیپٹ: ایجنٹ معمول کے سوالات کو یا تو متعین اصول کے مطابق یا ریپو میں پہلے سے درج سابق فیصلوں کی روشنی میں طے کرتے ہیں؛ جو ناقابلِ تنسیخ، مہنگا یا متنازع ہو وہ بانی کے پاس رہتا ہے۔ یہاں ہر ریکارڈ الٹنے کے قابل اور قابلِ جانچ ہے۔
273 فیصلہ ریکارڈز · 277 درج سابقے
حیثیت کے مطابق: طے شدہ (236) · بانی کے پاس (36) · recommended (1)
Should invention browser records live in a new class or the existing ideas wing? dec-20261108-001 · 2026-11-08
طے شدہ · precedent · لین: inventions · اعتماد: 0.95 · الٹنے کے قابل
فیصلہ: apply precedent D-001@inventions-decisions (decision-record): Use the existing `ideas` wing
Question overlaps precedent D-001@inventions-decisions recorded in docs/inventions/decisions.md. Conservative application; founder audits the sample.
سابقہ: D-001@inventions-decisions (docs/inventions/decisions.md)
Can we mark these 400 records reviewed and published? dec-20261108-002 · 2026-11-08
بانی کے پاس · policy-gate · لین: poc · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: review-status escalation is founder/pipeline only
Which shelf label for a 14th-century astronomical instrument treatise? dec-20261108-003 · 2026-11-08
طے شدہ · safe-default · لین: poc · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Canon decision process for the 286 undecided canon books: should the founder decide each book individually, or approve batches of decision records prepared by an agent from the ledger's checks? dec-20261108-004 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Content-Signal flip: add 'Content-Signal: search=yes, ai-input=yes, ai-train=no' as one line in site/public/robots.txt — the data is already CC0 so the flip grants nothing the licence does not; reversible. dec-20261108-005 · 2026-11-08
بانی کے پاس · policy-gate · لین: coverage-programme · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, headers, infra, CI) are founder-only
Corpus access for Phases 3, 4, 5, 7, 8: the founder runs the rijal name-match and returns counts; for Nadwi and DİA decide whether to purchase, scan or use the free index. dec-20261108-006 · 2026-11-08
بانی کے پاس · policy-gate · لین: coverage-programme · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: money is never autonomous
Lane registration for the coverage programme: register each phase as a sub-lane, or one programme lane with phase queues — the plan proposes the second. dec-20261108-007 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Muhaddithat scope and faith configuration for Phase 3: Nadwi's list includes women from several traditions — include the full list with labelled perspectives, or Sunni-only with the rest held back? dec-20261108-008 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Edge schema for rijal Phase 4: confirm the studied_under and transmitted-to predicates, or propose new ones through schema/predicates.yaml. dec-20261108-010 · 2026-11-08
بانی کے پاس · policy-gate · لین: coverage-programme · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, headers, infra, CI) are founder-only
Rijal inclusion rule for Phase 4: Tahdhib al-Kamal includes narrators from non-Sunni traditions — include them with labels following the Shia-mapping ruling, or exclude explicit non-Sunni narrators as the harvest rule does for persons? dec-20261108-011 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Commission the Sahaba lane: a separate harvest commission on the modern-figures machinery (al-Isaba / Usd al-Ghaba) sequenced after Phase 1 — commissioned only on the founder's explicit word. dec-20261108-012 · 2026-11-08
بانی کے پاس · policy-gate · لین: coverage-programme · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: commissions, money and external commitments are founder-only
Render-QA crawl for Phase 15: authorise a read-only crawl of the built HTML run locally — building locally is a check, not a deploy. dec-20261108-013 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Wikidata debt measurement for Phase 13: confirm which flag field marks the second-source debt so the 4,199 estimate can be checked against the 5,604 flagged files. dec-20261108-014 · 2026-11-08
طے شدہ · rule · لین: coverage-programme · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: children's content follows the children's wing rulebook: flagged_for_kids marks eligibility, sensitive topics get child_review, and leveled sections are sprouts/explorers/scholars inside the prose file
Answered by a standing repository rule (rule-kids-wing): docs/education + E5 leveled batches (founder-approved leveled-200 list).
Which flag field marks the ~4,199 records carrying second-source debt so the estimate can be checked against the 5,604 flagged files? dec-20261108-015 · 2026-11-08
recommended · research · لین: coverage-programme · اعتماد: 0.8 · الٹنے کے قابل
فیصلہ: There is no single flag field — measure the debt across three channels: needs_source flags (281 files), provenance notes saying 'not verified against a second source', and records citing only source/wikidata.
The question assumes one flag field exists; inspection shows the debt marker is distributed. A single-field count would under-report. The 4,199 vs 5,604 gap is explained by which channels are counted.
Policy question for the founder (one-time — sets config, not a per-item ruling): what false-discovery-rate target should the judge's conformal calibration use — alpha 0.02 (fewer autonomous settlements, more unknowns) or alpha 0.05 (more settlements, more risk of a false claim about a real person)? Global or per decision class? Note the asymmetry: a false settlement publishes a false claim; a false unknown publishes an honest gap — the classes are not symmetric. dec-20261108-016 · 2026-11-08
طے شدہ · safe-default · لین: poc · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Wikidata debt measurement for Phase 13: confirm which flag field marks the second-source debt so the 4,199 estimate can be checked against the 5,604 flagged files. dec-20261108-017 · 2026-11-08
طے شدہ · override · لین: poc · اعتماد: 1 · الٹنے کے قابل
فیصلہ: There is no single flag field — measure the debt across three channels: needs_source flags (281 files), provenance notes saying 'not verified against a second source', and records citing only source/wikidata.
dec-014 answered a Wikidata-debt question with the children's-wing rulebook because the pattern 'kid' matched inside 'wikidata'. The pattern was fixed (word boundaries) but the record was never overridden — it stayed live precedent authority and the benchmark certified the wrong match (Fable review F1). This override reverses it; the correct answer is the research-tier measurement from dec-015.
Can chess itself, or one of its variants reported in Muslim sources, enter the browser as a Sunni Muslim invention? dec-20261108-018 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: WCAG 2.2 AA: alt text required on every image, pages readable without JavaScript, islands lazy-load
Answered by a standing repository rule (rule-accessibility): AGENTS.md (accessibility rule).
Can Rufayda's tent be published although the sources disagree on what the first hospital in Islam was? dec-20261108-019 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: conflicting dates, names or readings are preserved, not resolved silently: record every reading attributed to its source, mark the field debated where the sources disagree, and follow the cited authority the record names
Answered by a standing repository rule (rule-conflict-preserve): commit history: Tansen, al-Yaqubi, wave-189/191 date conflicts 'recorded both ways, attributed'.
Can the broad madrasa/ijaza lead become a precise institutional contribution without repeating a first-university story? dec-20261108-020 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Can the existing double-acting suction-pump lead be completed with an explained mechanism, a licensed actual image and a distinction between the medieval design and modern reconstruction? dec-20261108-021 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Can the Fakhri sextant, a separate mural quadrant, or the *shamila* be promoted to invention records? dec-20261108-022 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Can the hand-washing automaton lead be completed without turning a surviving drawing into a surviving or certainly constructed machine? dec-20261108-023 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Can the historical development of Sufi khalwa or Nasta'liq script be attributed to an eligible Sunni Muslim maker? dec-20261108-024 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Can the intermittent gearing attributed to Ibn Khalaf al-Muradi be promoted to an invention record from the sources currently available? dec-20261108-025 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
D-017 kept the waqf as a lead, together with a comparison to the English trust. Who is the maker, what is claimed, and does the trust comparison stand? dec-20261108-026 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do Rufayda's medical tent, al-Walid's Damascus facility and Harun al-Rashid's Baghdad bimaristan establish a single early hospital invention? dec-20261108-027 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do the 15 queue leads derived from al-Farghani's astronomy concepts establish 15 inventions? dec-20261108-028 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do the broad kohl-cosmetic and miswak leads identify inventions made by eligible Sunni Muslim makers? dec-20261108-029 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do the catgut-suture and surgical-instrument leads establish a specific invention by an eligible Sunni maker? dec-20261108-030 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do the *kadhib*, *luzum*, *qadar*, *ri'aya* or Yawm Arafa concepts describe inventions or transmissions for this browser? dec-20261108-031 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: lanes process every queue item in order — nobody is skipped; what cannot be built is logged unidentifiable or held with its reason, never passed over silently
Answered by a standing repository rule (rule-nothing-skipped): Method page: 'lanes and waves with nobody skipped' + modern-figures §2 no-skip rule.
Do the linked early-chronology museum records identify an invention lead? dec-20261108-032 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: dates that are impossible on their face — born/died outside the person's documented activity window or contradicted by their own sourced events — are stripped entirely: the record carries no born/died blocks rather than wrong ones; never repaired by invention
Answered by a standing repository rule (rule-impossible-dates): R1 rulings 1 Nov 2026: Hadji Kamlon's queue dates 1890–1920 impossible vs documented 1948–55 rebellion — record carries no born/died blocks.
Do the mosque gnomon/sundial and Umayyad/Fatimid court-clock leads establish an original Muslim timekeeping invention? dec-20261108-033 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Do the screened sources establish a qualifying Muslim-origin claim for the Sistan vertical-axis windmill or a specific Muslim improvement to the saqiya? dec-20261108-034 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only with its cited basis, never inferred; absent a basis, omit the field or mark it unknown — contested cases use labelled perspectives, Sunni first
Answered by a standing repository rule (rule-faith-basis): AGENTS.md + docs/scope.md.
Do the two passage records and five work records identify inventions or transmissions within the browser's scope? dec-20261108-035 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does al-Umari's account establish a Muslim invention of determining the qibla by stars, winds, mountains or rivers? dec-20261108-036 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does another bimaristan add a distinct documented contribution rather than inflate the count with a building? dec-20261108-037 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does Mu'awiya's organization of the barid qualify as an original invention or a specific Muslim improvement? dec-20261108-038 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the account of Rufayda al-Aslamiyya's tent establish an invention or medical institution attributable to her? dec-20261108-039 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the broad turban lead identify a Sunni Muslim invention? dec-20261108-040 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the current evidence establish that al-Razi first distinguished smallpox from measles as an eligible Sunni Muslim maker? dec-20261108-041 · 2026-11-08
طے شدہ · precedent · لین: seed · اعتماد: 0.8 · الٹنے کے قابل
فیصلہ: apply precedent D-077 (decision-record): Al-Razi's documented smallpox/measles comparison as invention37 (6 November2026)
Question overlaps precedent D-077 recorded in docs/inventions/decisions.md. Conservative application; founder audits the sample.
سابقہ: D-077 (docs/inventions/decisions.md)
Does the current lead identify a specific Muslim adaptation of Byzantine Greek fire, with an eligible maker? dec-20261108-042 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the current waqf lead support a comparison with the English trust or a single invention attribution? dec-20261108-043 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the early minaret lead establish that Maslama ibn Mukhallad invented the minaret as an architectural form? dec-20261108-044 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the fountain-pen lead attributed to al-Mu'izz identify a qualifying Muslim maker? dec-20261108-045 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the game's early passage into Muslim societies establish a qualifying Muslim transmission record? dec-20261108-046 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only on two independent sources attesting it for the person himself; office, headship, descent, movement membership, infobox-only values, teacher lineages and aggregator rows never count; where sources say only 'Shia' and context supports a branch, the founder's mapping gives shia-twelver documented with the mapping stated — otherwise unknown
Answered by a standing repository rule (rule-faith-two-source): Method page doctrine + founder Shia-mapping rule 24 Oct 2026 + founder re-review 1 Nov 2026 ('strict two-source rule restored; branch-only attestation settles nothing').
Does the generic census hit for Ibn Taymiyya identify a specific invention lead? dec-20261108-047 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the generic timekeeping census hit for Kamal Khujandi identify a relevant invention contribution? dec-20261108-048 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the generic timekeeping census hit for Ulubatlı Hasan identify a specific invention contribution? dec-20261108-049 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the generic timekeeping event-census hit for the Alpujarras Rebellion identify a specific invention? dec-20261108-050 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the hisba and market-inspector lead identify a Muslim invention? dec-20261108-051 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the planning/census work require a release version bump? dec-20261108-052 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: every change appends its entry to CANARY.md in the same commit under the current version block; breaking → major, new feature/page/class/endpoint → minor, fixes → patch
Answered by a standing repository rule (rule-canary): AGENTS.md (canary release notes).
Does the primary text establish the hollow cataract needle as an eligible Sunni Muslim invention? dec-20261108-053 · 2026-11-08
طے شدہ · precedent · لین: seed · اعتماد: 0.8 · الٹنے کے قابل
فیصلہ: apply precedent D-110 (decision-record): Invention69: Ammar al-Mawsili's described hollow cataract needle
Question overlaps precedent D-110 recorded in docs/inventions/decisions.md. Conservative application; founder audits the sample.
سابقہ: D-110 (docs/inventions/decisions.md)
Does the source evidence establish a specific change to the barid system, and is Mu'awiya's maker eligibility ready for promotion? dec-20261108-054 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does the tiraz lead establish a Muslim invention of textile production? dec-20261108-055 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Equations were solved in Egypt, Babylon, Greece and India long before al-Khwarizmi, and modern historians doubt that his book was the first on al-jabr. What can the record say he made, and how is the priority question shown? dec-20261108-056 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: where there is doubt the classification is unknown/unclassified — doubt is not establishment; a name-basis or single weak attestation never settles a sensitive field; record the contested readings preserved in the note
Answered by a standing repository rule (rule-doubt-unknown): founder doubt rule, 24 Oct 2026 (docs/agents/lanes/modern-figures.md §3; revert commit 'founder re-review' 1 Nov 2026: 'the doubt rule is absolute').
How should a system that existed under the Persians and Byzantines, and whose first Islamic organiser is reported in two ways, be recorded? dec-20261108-057 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
How should al-Maqrizi's statement that al-Walid ibn Abd al-Malik was "the first to build a maristan in Islam" be recorded, when Rufayda's tent is already published as a candidate first and the modern sources place the house in Damascus without a primary citation? dec-20261108-058 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: any source may be read and its facts enter data/ with a citation; Wikipedia and Wikidata are pointers for entity facts taken from a recorded revision — never sources for prose or a work's content; media only from the licence whitelist
Answered by a standing repository rule (rule-source-admission): docs/sources-policy.md + docs/agents/source-families.md.
How should invention dates preserve a Hijri date when the Phase 1 example uses `calendar: hijri`? dec-20261108-059 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: never invent a date: EDTF + Hijri string + calendar + precision; conversions marked derived; one source is enough, a doubt is flagged honestly
Answered by a standing repository rule (rule-dates): AGENTS.md (dates rule).
How should the census count the prompt's scientist/person pool? dec-20261108-060 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Is the Baghdad hospital a new invention, an improvement on Jundishapur or the transfer of an earlier model? dec-20261108-061 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Light passing through a small hole into a dark room is an everyday sight that Ibn al-Haytham himself calls familiar, and popular accounts call him the inventor of the camera. What can the record claim? dec-20261108-062 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Popular accounts credit al-Khwarizmi with "inventing the algorithm" or even the numerals. What can the record claim? dec-20261108-063 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Should Phase 1 also add `/api/v1/inventions/timeline.json`, which appears in the operative prompt's later browser requirements? dec-20261108-064 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: every feature is graph + API + search: records stay in the one graph, the class is exported in the static API, pages are search-indexed, and the feature is registered in docs/site/browsers.md, the plan index, the roadmap, docs/agents/lanes.md and docs/status.md — label anything submission-impacting
Answered by a standing repository rule (rule-feature-graph-api-search): AGENTS.md (every feature is graph, API and search) + docs/ai-change-gates.md §3.
Should the browser wait for a full first batch, or ship as soon as one record meets every gate? dec-20261108-065 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
The hisba is an office, not a device, and the sources place the first market officers in the Prophet's time; orientalists derived it from the Greek agoranomos or the Byzantine prefect. Who is the maker, and what is claimed? dec-20261108-066 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
The tiraz band on royal cloth is older than Islam, and the primary report about Abd al-Malik ibn Marwan concerns writing at the head of papyrus rolls, not cloth. How should the record frame the invention, its maker and its date? dec-20261108-067 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: never invent a date: EDTF + Hijri string + calendar + precision; conversions marked derived; one source is enough, a doubt is flagged honestly
Answered by a standing repository rule (rule-dates): AGENTS.md (dates rule).
What specific al-Jazari contribution can be attributed without converting popular robot and hidden-mechanism summaries into unsupported facts? dec-20261108-068 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Which existing wing should invention records use? dec-20261108-069 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Which product release version should record the Phase 1 class and API after `origin/main` advances? dec-20261108-070 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Who is the maker of the first minarets when the sources name the governor who ordered them but not the builders? dec-20261108-071 · 2026-11-08
طے شدہ · safe-default · لین: seed · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
World maps and the seven climes are older than al-Idrisi, and his preface gives Roger II, a Christian king, the leading part in the inquiry. What can the record say was made, by whom, and where? dec-20261108-072 · 2026-11-08
طے شدہ · rule · لین: seed · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only on two independent sources attesting it for the person himself; office, headship, descent, movement membership, infobox-only values, teacher lineages and aggregator rows never count; where sources say only 'Shia' and context supports a branch, the founder's mapping gives shia-twelver documented with the mapping stated — otherwise unknown
Answered by a standing repository rule (rule-faith-two-source): Method page doctrine + founder Shia-mapping rule 24 Oct 2026 + founder re-review 1 Nov 2026 ('strict two-source rule restored; branch-only attestation settles nothing').
My feature adds /api/v1/inventions/timeline.json — update site/public/openapi/v1.json and llms.txt in the same commit, or ship the endpoint and docs in a follow-up? dec-20261108-073 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: every feature is graph + API + search: records stay in the one graph, the class is exported in the static API, pages are search-indexed, and the feature is registered in docs/site/browsers.md, the plan index, the roadmap, docs/agents/lanes.md and docs/status.md — label anything submission-impacting
Answered by a standing repository rule (rule-feature-graph-api-search): AGENTS.md (every feature is graph, API and search) + docs/ai-change-gates.md §3.
My new Arabic prose has three parenthetical asides — fix the parentheses isolation in the same commit as the prose, or land the prose and fix punctuation in a follow-up? dec-20261108-074 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
My rules.yaml edit dropped benchmark agreement from 92/92 to 91/92 — revert the rule change, or update the benchmark expectation if the new behaviour is correct? dec-20261108-075 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A schema change merged an hour ago and 40 content entities are queued — deploy now for the schema fix or batch everything into tomorrow's scheduled deploy? dec-20261108-076 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
docs/plan/06-roadmap.md says Phase 2 starts after Phase 1 sign-off but AGENTS.md describes lanes running in parallel — which document governs my sequencing decision? dec-20261108-077 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: conflicting authorities require founder resolution — no rule or precedent can outrank another
My change only edits docs/decisions/rules.yaml comments — does it still need a CANARY.md entry and a full validate run before commit? dec-20261108-078 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: every change appends its entry to CANARY.md in the same commit under the current version block; breaking → major, new feature/page/class/endpoint → minor, fixes → patch
Answered by a standing repository rule (rule-canary): AGENTS.md (canary release notes).
check-extremism flags a URL in my new source record — rewrite the citation to describe the publisher neutrally, or request a founder exception entry? dec-20261108-079 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
docs/coverage.md has a merge conflict after merging main — hand-resolve the YAML conflict or delete the file and regenerate it with the coverage script? dec-20261108-080 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: destructive operations are never autonomous — negation tricks must not disarm this gate
My UI change adds a new string — ship en.yaml and ur.yaml translations in the same commit as required, or land the Arabic string and add translations in the next batch? dec-20261108-081 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.8 · الٹنے کے قابل
فیصلہ: ship the change inside the current layout/component pattern — a brief that requires deviation becomes a labelled proposal logged for review, never a silent redesign
Answered by a standing repository rule (rule-ui-no-deviation): AGENTS.md UI constraint + Muse Q2 domain table (no-big-deviation rule).
My batch run minted dec-20261108-073 but a record with that id already exists on origin/main from another session — renumber my record, or keep the id since the questions differ? dec-20261108-082 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Another agent session registered the rijaal lane yesterday and I have been working assumptions under a different lane name — register a sub-lane, adopt their lane, or ask the founder? dec-20261108-083 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
origin/main advanced 12 commits while my branch was open, including edits to the same lane progress file — rebase onto main, merge main into my branch, or restart the branch? dec-20261108-084 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: deduplicate: verify identity against the sources first; if it is one entity, merge into the richer record and keep the other id under aliases — never delete blindly; if they are distinct, keep both
Answered by a standing repository rule (rule-duplicate-alias): AGENTS.md (reuse before create; a duplicate becomes an alias of the richer entity).
Two sessions edited the same entity file in parallel worktrees and both branches touched data/person/ibn-sina.yaml — merge both versions field-by-field or keep my branch's version? dec-20261108-085 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: deduplicate: verify identity against the sources first; if it is one entity, merge into the richer record and keep the other id under aliases — never delete blindly; if they are distinct, keep both
Answered by a standing repository rule (rule-duplicate-alias): AGENTS.md (reuse before create; a duplicate becomes an alias of the richer entity).
My entity batch has 31 records ready — split into two PRs of ~15 each to respect the small-PR convention, or ship all 31 in one PR since they are one logical batch? dec-20261108-086 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
npm run validate fails on a pre-existing error in an entity file I did not touch — do I fix it in my PR, file it separately, or proceed anyway? dec-20261108-087 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
My AI-generated prose cites two defined sources and passes validation — mark it status.review: published now, or hold it as draft until a human reads it? dec-20261108-088 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: review-status escalation is founder/pipeline only
I need to model a new relation type for the inventions browser — do I register the predicate in schema/predicates.yaml first and write entities second, or create entities and add the predicate after? dec-20261108-089 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
My worktree branch is 40 commits behind main and my lane files were touched upstream — merge main into my branch now, or finish the current batch first and merge at the end? dec-20261108-090 · 2026-11-08
طے شدہ · rule · لین: process-pilot · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: deduplicate: verify identity against the sources first; if it is one entity, merge into the richer record and keep the other id under aliases — never delete blindly; if they are distinct, keep both
Answered by a standing repository rule (rule-duplicate-alias): AGENTS.md (reuse before create; a duplicate becomes an alias of the richer entity).
A vitest case fails on my machine but passes in CI and on a clean checkout — investigate the local environment, mark it flaky, or ignore it since CI is green? dec-20261108-091 · 2026-11-08
طے شدہ · safe-default · لین: process-pilot · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
This batch fixes a date-parsing bug and adds one new predicate — do I bump the site version as patch, minor, or leave it for the next feature batch? dec-20261108-092 · 2026-11-08
بانی کے پاس · policy-gate · لین: process-pilot · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
A building photo shows the site today but the entity is historical — use it as contextual image with a caption, or find period imagery only? dec-20261108-093 · 2026-11-08
طے شدہ · rule · لین: imagery · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: media only from schema/enums/licenses.yaml; third-party content keeps its own licence; no figurative depictions of prophets or Companions; no AI-generated portraits
Answered by a standing repository rule (rule-media-license): AGENTS.md (licences rule).
A canon book decision affects a filed DPGA/UNESCO submission — flag it submission-impacting and escalate, or treat as routine canon work? dec-20261108-094 · 2026-11-08
طے شدہ · safe-default · لین: submission · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A historical topic is valid but the only detailed source carries modern extremist framing — rewrite from an academic angle per the hard-block rule, or skip the topic? dec-20261108-095 · 2026-11-08
بانی کے پاس · policy-gate · لین: content-safety · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: content-safety hard block — always founder
A map PMTiles file needs regeneration after new place entities — regenerate in the same commit or a follow-up data commit? dec-20261108-096 · 2026-11-08
طے شدہ · safe-default · لین: maps · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A modern politician's museum entry would need evaluative language (hero/villain framing) — describe neutrally with story_roles or decline the entry? dec-20261108-097 · 2026-11-08
طے شدہ · safe-default · لین: voice · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A new media license appears in a source I want to cite — do I add it to schema/enums/licenses.yaml first, or cite and flag the gap? dec-20261108-098 · 2026-11-08
بانی کے پاس · policy-gate · لین: licensing · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
A new UI string has Arabic and English but Urdu translation is not ready — ship ar+en now and add ur next batch, or hold the whole change? dec-20261108-099 · 2026-11-08
طے شدہ · safe-default · لین: translation · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A person's conversion to Islam is documented by exactly one source — record the conversion event with one source flagged, or hold for a second? dec-20261108-100 · 2026-11-08
طے شدہ · precedent · لین: faith · اعتماد: 0.8 · الٹنے کے قابل
فیصلہ: apply precedent D-061 (decision-record): Hold the proposed post-conversion al-Samawal calculation
Question overlaps precedent D-061 recorded in docs/inventions/decisions.md. Conservative application; founder audits the sample.
سابقہ: D-061 (docs/inventions/decisions.md)
A person's madhhab is not stated but both his teachers were Shafi'i — record shafi'i with a derived flag, or leave unknown? dec-20261108-101 · 2026-11-08
طے شدہ · safe-default · لین: faith · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
A quotation I need runs 60 words — cite it verbatim anyway, truncate with ellipsis, or paraphrase in our own words? dec-20261108-102 · 2026-11-08
طے شدہ · rule · لین: quoting · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: no verbatim text beyond a 40-word cited quotation enters git; rewrite in the museum's own words with a citation
Answered by a standing repository rule (rule-40-word-quote): docs/sources-policy.md.
A revert is needed for a bad merge — does the rollback get its own CANARY entry naming what was reverted? dec-20261108-103 · 2026-11-08
طے شدہ · rule · لین: workflow · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: every change appends its entry to CANARY.md in the same commit under the current version block; breaking → major, new feature/page/class/endpoint → minor, fixes → patch
Answered by a standing repository rule (rule-canary): AGENTS.md (canary release notes).
A scholarly source claims consensus on a contested point without citing who agrees — soften to attributed claim or reject the source? dec-20261108-104 · 2026-11-08
طے شدہ · rule · لین: admission · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: every exclusion is a proposal awaiting the founder's decision, never a final exclusion: relabel the row 'held for the founder', keep the record, move source-less items to the leads list — nothing is dropped silently
Answered by a standing repository rule (rule-exclusions-are-proposals): founder rule of 7 Nov 2026 (docs/modern-islam progress tables; commit history).
A source I cited last month published a correction — update my citation to the corrected revision or keep the original with a note? dec-20261108-105 · 2026-11-08
طے شدہ · rule · لین: corrections · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: corrections are made openly: fix the record, cite what the correction rests on, append a CANARY entry — errors are corrected when found, never hidden
Answered by a standing repository rule (rule-corrections-open): AGENTS.md + Method page: 'corrections in the open'.
A stub entity has almost no data — enrich it in place to full record, or leave the stub and create the full record under a new id? dec-20261108-106 · 2026-11-08
طے شدہ · rule · لین: entities · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: stubs are replaced in place with fully sourced records — rebuild at the same slug, repoint inbound edges and wikilinks, keep the richer version; a stub is never silently skipped while sources exist
Answered by a standing repository rule (rule-stub-replace): founder rulings review 3 Nov 2026: 'REPLACE on all stubs' (wave 185).
A tension card has only one cited voice so far — publish with the single voice flagged, or wait for a second perspective? dec-20261108-107 · 2026-11-08
طے شدہ · rule · لین: perspectives · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: a tension card carries at least two cited voices with labelled perspectives and no verdict
Answered by a standing repository rule (rule-tension-card): AGENTS.md (voice and scope).
A useful fact comes only from a personal blog with no author date — admissible as a source for data/ or prose? dec-20261108-108 · 2026-11-08
طے شدہ · safe-default · لین: admission · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
An entity has a data record but no prose file yet — publish the stub page now or wait until prose exists? dec-20261108-109 · 2026-11-08
طے شدہ · rule · لین: prose · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: stubs are replaced in place with fully sourced records — rebuild at the same slug, repoint inbound edges and wikilinks, keep the richer version; a stub is never silently skipped while sources exist
Answered by a standing repository rule (rule-stub-replace): founder rulings review 3 Nov 2026: 'REPLACE on all stubs' (wave 185).
An event has only a Hijri date (no Gregorian conversion source) — store hijri-only with precision, or derive the CE range marked derived? dec-20261108-110 · 2026-11-08
طے شدہ · safe-default · لین: chronology · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
An existing entity is missing label.ur — backfill it inside my unrelated PR or file it as separate i18n debt? dec-20261108-111 · 2026-11-08
طے شدہ · safe-default · لین: translation · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
An image is CC BY-NC (non-commercial) — usable on the site given the museum is free, or does NC disqualify it? dec-20261108-112 · 2026-11-08
طے شدہ · safe-default · لین: licensing · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
data/_stubs.yaml still lists an entity I just wrote fully — remove the stub entry in the same commit or leave it for the next cleanup pass? dec-20261108-113 · 2026-11-08
طے شدہ · rule · لین: data-hygiene · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: stubs are replaced in place with fully sourced records — rebuild at the same slug, repoint inbound edges and wikilinks, keep the richer version; a stub is never silently skipped while sources exist
Answered by a standing repository rule (rule-stub-replace): founder rulings review 3 Nov 2026: 'REPLACE on all stubs' (wave 185).
docs/coverage.md is stale after my entity batch — regenerate it with the coverage script or hand-edit the counts? dec-20261108-114 · 2026-11-08
طے شدہ · safe-default · لین: docs · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
docs/status.md ownership table conflicts with my lane assignment in docs/agents/lanes.md — ask the founder or follow lanes.md? dec-20261108-115 · 2026-11-08
طے شدہ · safe-default · لین: lanes · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
docs/status.md says lane X is paused but AGENTS.md lanes.md says active — which file do I trust for lane state? dec-20261108-116 · 2026-11-08
طے شدہ · safe-default · لین: lanes · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
I added a field to an API export — update site/public/openapi/v1.json and llms.txt in the same commit or a docs follow-up? dec-20261108-117 · 2026-11-08
طے شدہ · rule · لین: api · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: every feature is graph + API + search: records stay in the one graph, the class is exported in the static API, pages are search-indexed, and the feature is registered in docs/site/browsers.md, the plan index, the roadmap, docs/agents/lanes.md and docs/status.md — label anything submission-impacting
Answered by a standing repository rule (rule-feature-graph-api-search): AGENTS.md (every feature is graph, API and search) + docs/ai-change-gates.md §3.
I added a Worker route — do I update wrangler.production.toml run_worker_first and the deploy.yml removal lines in the same commit? dec-20261108-118 · 2026-11-08
بانی کے پاس · policy-gate · لین: infra · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
I need a predicate that does not exist in schema/predicates.yaml — propose it through the governance docket or use a prose mention instead? dec-20261108-119 · 2026-11-08
بانی کے پاس · policy-gate · لین: schema · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
main is broken by someone else's merge (build fails) — hotfix it immediately in my PR or report and let the owner fix? dec-20261108-120 · 2026-11-08
طے شدہ · rule · لین: workflow · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: fix it in its own small commit: reproduce first, root-cause it, run validate + tests (+ build for site changes), append a CANARY entry — never hide a failure behind a guard or a widened timeout
Answered by a standing repository rule (rule-bug-fix): docs/agents/workflow.md + docs/ai-change-gates.md §5.
My change adds files under a new data/ folder — which lists must I update for the R2 sync (R2_DATA_DIRS, publish.mjs, run_worker_first, deploy.yml)? dec-20261108-121 · 2026-11-08
بانی کے پاس · policy-gate · لین: infra · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
My change edits a file under site/public/shell/ which rewrites all ~164k pages — label the PR rewrites-all-pages and batch it, or find a way to avoid the rewrite? dec-20261108-122 · 2026-11-08
طے شدہ · safe-default · لین: deploy · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
My change only touches .github/workflows/ci.yml — does it still need the full validate + test + build gate chain before push? dec-20261108-123 · 2026-11-08
بانی کے پاس · policy-gate · لین: ci · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
My lane progress file conflicts with another lane's edits on merge — keep both sections, take mine, or regenerate from the lane logs? dec-20261108-124 · 2026-11-08
طے شدہ · safe-default · لین: workflow · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
My PR is green on validate and test — should I merge it myself or wait for a reviewer? dec-20261108-125 · 2026-11-08
طے شدہ · safe-default · لین: workflow · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
npm test fails in an unrelated suite that was already failing on main — proceed with my change or fix that suite first? dec-20261108-126 · 2026-11-08
طے شدہ · safe-default · لین: testing · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Pagefind index is stale after yesterday's content merges — rebuild the index now, at next deploy, or on a schedule? dec-20261108-127 · 2026-11-08
بانی کے پاس · policy-gate · لین: search · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
The Arabic label exists but English summary is machine-translated and unreviewed — publish with provenance, or hold until human-reviewed? dec-20261108-128 · 2026-11-08
طے شدہ · safe-default · لین: translation · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Two independent sources give different death years for the same person — publish one year with a note, or record both readings as debated? dec-20261108-129 · 2026-11-08
طے شدہ · rule · لین: chronology · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: deduplicate: verify identity against the sources first; if it is one entity, merge into the richer record and keep the other id under aliases — never delete blindly; if they are distinct, keep both
Answered by a standing repository rule (rule-duplicate-alias): AGENTS.md (reuse before create; a duplicate becomes an alias of the richer entity).
Two open PRs edit the same entity file in different fields — merge them sequentially with a 3-way merge each time, or combine into one PR? dec-20261108-130 · 2026-11-08
طے شدہ · rule · لین: workflow · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: fetch origin/main first and review what arrived; three-way merge keeping both sides' intent; never force-push, never overwrite or revert another session's work; explicit paths only, never git add -A
Answered by a standing repository rule (rule-merge): docs/agents/workflow.md §4 + docs/ai-change-gates.md §4.
Two works share the same title in different languages — merge into one record with language variants, or keep separate records per tradition? dec-20261108-131 · 2026-11-08
طے شدہ · rule · لین: entities · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: deduplicate: verify identity against the sources first; if it is one entity, merge into the richer record and keep the other id under aliases — never delete blindly; if they are distinct, keep both
Answered by a standing repository rule (rule-duplicate-alias): AGENTS.md (reuse before create; a duplicate becomes an alias of the richer entity).
Can we mark these 400 records reviewed and published? dec-20261108-132 · 2026-11-08
طے شدہ · override · لین: poc · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Do not batch-mark the ~400 flagged records reviewed/published — each requires individual review before publication.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-002 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Canon decision process for the 286 undecided canon books: should the founder decide each book individually, or approve batches of decision records prepared by an agent from the ledger's checks? dec-20261108-133 · 2026-11-08
طے شدہ · override · لین: coverage-programme · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The founder does not decide each canon book individually — agent-prepared batches of 25 proceed under delegation, with the founder approving each batch (not each book) before register entries are written.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-004 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08 Clarified 2026-11-08 (Claude audit #3): 'approve batches' means founder batch-level sign-off, consistent with authenticity-policy-proposal.md §5 and the phase-02 queue header.
Policy question for the founder (one-time — sets config, not a per-item ruling): what false-discovery-rate target should the judge's conformal calibration use — alpha 0.02 (fewer autonomous settlements, more unknowns) or alpha 0.05 (more settlements, more risk of a false claim about a real person)? Global or per decision class? Note the asymmetry: a false settlement publishes a false claim; a false unknown publishes an honest gap — the classes are not symmetric. dec-20261108-134 · 2026-11-08
طے شدہ · override · لین: poc · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Use the stricter false-discovery-rate target alpha 0.02 for judge conformal calibration, not 0.05.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-016 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08 Verified 2026-11-08: alpha 0.02 applies GLOBALLY (founder confirmed — not per-class).
Can the historical development of Sufi khalwa or Nasta'liq script be attributed to an eligible Sunni Muslim maker? dec-20261108-135 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The historical development of Sufi khalwa or Nasta'liq script cannot be attributed to an eligible Sunni Muslim maker — the lead does not qualify.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-024 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Do Rufayda's medical tent, al-Walid's Damascus facility and Harun al-Rashid's Baghdad bimaristan establish a single early hospital invention? dec-20261108-136 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Rufayda's medical tent, al-Walid's Damascus facility and Harun al-Rashid's Baghdad bimaristan do NOT establish a single hospital invention.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-027 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Do the 15 queue leads derived from al-Farghani's astronomy concepts establish 15 inventions? dec-20261108-137 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The 15 queue leads derived from al-Farghani's astronomy concepts do NOT establish 15 inventions.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-028 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Do the broad kohl-cosmetic and miswak leads identify inventions made by eligible Sunni Muslim makers? dec-20261108-138 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The broad kohl-cosmetic and miswak leads do NOT identify inventions made by eligible Sunni Muslim makers.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-029 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Do the mosque gnomon/sundial and Umayyad/Fatimid court-clock leads establish an original Muslim timekeeping invention? dec-20261108-139 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The mosque gnomon/sundial and Umayyad/Fatimid court-clock leads do NOT establish an original Muslim timekeeping invention.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-033 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does al-Umari's account establish a Muslim invention of determining the qibla by stars, winds, mountains or rivers? dec-20261108-140 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Al-Umari's account does NOT establish a Muslim invention of determining the qibla by stars, winds, mountains or rivers.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-036 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the broad turban lead identify a Sunni Muslim invention? dec-20261108-141 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The broad turban lead does NOT identify a Sunni Muslim invention.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-040 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the early minaret lead establish that Maslama ibn Mukhallad invented the minaret as an architectural form? dec-20261108-142 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The early minaret lead does NOT establish that Maslama ibn Mukhallad invented the minaret as an architectural form.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-044 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the generic census hit for Ibn Taymiyya identify a specific invention lead? dec-20261108-143 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The generic census hit for Ibn Taymiyya does NOT identify a specific invention lead.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-047 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the generic timekeeping census hit for Kamal Khujandi identify a relevant invention contribution? dec-20261108-144 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The generic timekeeping census hit for Kamal Khujandi does NOT identify a relevant invention contribution.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-048 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the generic timekeeping census hit for Ulubatlı Hasan identify a specific invention contribution? dec-20261108-145 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The generic timekeeping census hit for Ulubatlı Hasan does NOT identify a specific invention contribution.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-049 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the generic timekeeping event-census hit for the Alpujarras Rebellion identify a specific invention? dec-20261108-146 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The generic timekeeping event-census hit for the Alpujarras Rebellion does NOT identify a specific invention.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-050 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Does the hisba and market-inspector lead identify a Muslim invention? dec-20261108-147 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The hisba and market-inspector lead does NOT identify a Muslim invention.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-051 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Light passing through a small hole into a dark room is an everyday sight that Ibn al-Haytham himself calls familiar, and popular accounts call him the inventor of the camera. What can the record claim? dec-20261108-148 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Ibn al-Haytham's record may NOT claim camera invention — credit the pinhole observation only.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-062 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Popular accounts credit al-Khwarizmi with "inventing the algorithm" or even the numerals. What can the record claim? dec-20261108-149 · 2026-11-08
طے شدہ · override · لین: seed · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Al-Khwarizmi's record may NOT claim 'inventing the algorithm' or the numerals — credit the algebra text only.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-063 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
My new Arabic prose has three parenthetical asides — fix the parentheses isolation in the same commit as the prose, or land the prose and fix punctuation in a follow-up? dec-20261108-150 · 2026-11-08
طے شدہ · override · لین: process-pilot · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Arabic parenthesis isolation is fixed in the SAME commit as the prose — no follow-up deferral.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-074 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
A schema change merged an hour ago and 40 content entities are queued — deploy now for the schema fix or batch everything into tomorrow's scheduled deploy? dec-20261108-151 · 2026-11-08
طے شدہ · override · لین: process-pilot · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Do NOT deploy immediately for the merged schema fix — batch the deploy with the next content deploy.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-076 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
My AI-generated prose cites two defined sources and passes validation — mark it status.review: published now, or hold it as draft until a human reads it? dec-20261108-152 · 2026-11-08
طے شدہ · override · لین: process-pilot · اعتماد: 1 · الٹنے کے قابل
فیصلہ: AI-generated prose citing defined sources does NOT get marked status.review: published without human read — it stays draft until reviewed.
The engine held this at safe-default/docket, but the founder's verdict on dec-20261108-088 was a substantive "no" — a ruling on the underlying question, not a rejection of the hold. This override enacts that ruling so the answer becomes citable precedent instead of a dead end. Source verdict note: batch verdict 2026-11-08
Two independent sources give different death years for the same person — publish one year with a note, or record both readings as debated? dec-20261108-153 · 2026-11-08
طے شدہ · override · لین: chronology · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Conflicting dates are preserved, not resolved silently: record every reading as a debated/attributed variant with its sources — never pick one year and bury the other.
dec-129 was generated before the conflict-preserve/source-family pattern tightening and cites rule-duplicate-alias — an entity-dedup rule answering a date-conflict question, the same false-positive class as the git-merge misfires. The current engine answers this question with rule-conflict-preserve. This override kills the stale record so it cannot teach the wrong ruling to future similar questions.
A historical topic is valid but the only detailed source carries modern extremist framing — rewrite from an academic angle per the hard-block rule, or skip the topic? dec-20261108-154 · 2026-11-08
طے شدہ · override · لین: content-safety · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Sources carrying modern extremist framing are blocked — not ingested. A valid historical topic is sourced from academic, art-historical or archaeological alternatives instead.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: founder ruled: do not ingest extremist-framing sources — block them entirely
A modern politician's museum entry would need evaluative language (hero/villain framing) — describe neutrally with story_roles or decline the entry? dec-20261108-155 · 2026-11-08
طے شدہ · override · لین: voice · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Modern politicians' museum entries use neutral description only — evaluative hero/villain framing is stripped, acts are described and judgements attributed.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: neutral only — strip evaluative language for modern politicians
A new UI string has Arabic and English but Urdu translation is not ready — ship ar+en now and add ur next batch, or hold the whole change? dec-20261108-156 · 2026-11-08
طے شدہ · override · لین: translation · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Do not ship incomplete Urdu translation files — Arabic, English and Urdu land in the same commit or not at all.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: do not ship incomplete Urdu translation files
A person's madhhab is not stated but both his teachers were Shafi'i — record shafi'i with a derived flag, or leave unknown? dec-20261108-157 · 2026-11-08
طے شدہ · override · لین: faith · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Madhhab is never inferred from a person's teachers — record only an explicitly cited affiliation; otherwise it stays unknown.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: no madhhab inference from teachers — must be explicitly cited
A useful fact comes only from a personal blog with no author date — admissible as a source for data/ or prose? dec-20261108-158 · 2026-11-08
طے شدہ · override · لین: admission · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Undated personal blogs are inadmissible as sources — a fact needs a datable, attributable source.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: undated personal blogs are inadmissible as sources
main is broken by someone else's merge (build fails) — hotfix it immediately in my PR or report and let the owner fix? dec-20261108-159 · 2026-11-08
طے شدہ · override · لین: workflow · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Do not hotfix a broken main inside a feature PR — report it separately and open a dedicated hotfix PR; feature PRs keep clean scopes.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: no hotfix of broken main inside feature PR — report separately, clean scopes
My PR is green on validate and test — should I merge it myself or wait for a reviewer? dec-20261108-160 · 2026-11-08
طے شدہ · override · لین: workflow · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Agents do not merge their own PRs — the founder merges. Green validate/test is necessary but not sufficient; merge authority stays with the founder.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: agents do not self-merge — founder merges
The Arabic label exists but English summary is machine-translated and unreviewed — publish with provenance, or hold until human-reviewed? dec-20261108-161 · 2026-11-08
طے شدہ · override · لین: translation · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Machine-translated summaries require human review before publishing — unreviewed MT prose stays draft.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: machine-translated summaries require human review before publishing
npm run validate fails on a pre-existing error in an entity file I did not touch — do I fix it in my PR, file it separately, or proceed anyway? dec-20261108-162 · 2026-11-08
طے شدہ · override · لین: process-pilot · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Do not carry a pre-existing validate failure inside a feature PR — file it separately; if main is broken, open a dedicated hotfix PR and rebase the feature branch onto it. The merge bar is 0 errors and PRs keep clean scopes.
The engine held/docketed this; the founder's wave-2 verdict was a substantive "no" — a ruling on the underlying question. This override enacts it so the answer becomes citable precedent. Source verdict note: batch verdict 2026-11-08
EI2 queue row AMIR KHAN (p. 466-467) contains a headword inside it — AMIR KHUSRAW — with no queue row of its own. Should a new record be created for Amir Khusraw from that embedded headword, or is it covered by an existing record? dec-20261108-163 · 2026-11-08
طے شدہ · rule · لین: ei2 · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: reuse before create: check existing records, aliases, stubs and wanted lists first; a duplicate becomes an alias of the richer entity; at most 25 entities per batch; Arabic label+summary required with en+ur in the same commit; every entity cites sources
Answered by a standing repository rule (rule-create-entity): AGENTS.md (graph in git + sources + language rules).
The Azhar academy book gives the Battle of Ayn Jalut as Friday 25 Ramadan 658 AH while the existing event entity gives Friday 15 Ramadan 658 — which Hijri date should the record carry, and how should the conflict be represented? dec-20261108-164 · 2026-11-08
طے شدہ · rule · لین: azhar-academy · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: fetch origin/main first and review what arrived; three-way merge keeping both sides' intent; never force-push, never overwrite or revert another session's work; explicit paths only, never git add -A
Answered by a standing repository rule (rule-merge): docs/agents/workflow.md §4 + docs/ai-change-gates.md §4.
Mustafa ibn Bayezid's record carries a contested claim about the Düzme question — how should a contested parentage/attribution claim be recorded: attributed debated reading, perspectives block, or omit pending resolution? dec-20261108-165 · 2026-11-08
طے شدہ · rule · لین: dawlat-uthmaniyya · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: contested topics use the perspectives block: Sunni position(s) first, then the academic view, then any other view — each cited and labelled, no verdict in the museum's voice
Answered by a standing repository rule (rule-perspectives): AGENTS.md (voice and scope).
EI2 queue seq 273 mixes the Gibb text of AMIR AL-MU'MININ with a BSOAS tail — which headword should the record be written under, AMIR AL-MU'MININ or the BSOAS article, or split into two records? dec-20261108-166 · 2026-11-08
طے شدہ · safe-default · لین: ei2 · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
The Azhar academy book's chapter on Ibtida' in religion overlaps the existing concept/bida entity — should the chapter's content extend concept/bida in place, or warrant a separate entity? dec-20261108-167 · 2026-11-08
طے شدہ · rule · لین: azhar-academy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: reuse before create: check existing records, aliases, stubs and wanted lists first; a duplicate becomes an alias of the richer entity; at most 25 entities per batch; Arabic label+summary required with en+ur in the same commit; every entity cites sources
Answered by a standing repository rule (rule-create-entity): AGENTS.md (graph in git + sources + language rules).
The khutab lane is blocked pending the founder's decision on the Sulayman entry and the Umar ibn Abd al-Aziz slug — which canonical ids should those entities use? dec-20261108-168 · 2026-11-08
بانی کے پاس · founder-reserved · لین: khutab-durus · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: question reserves this decision to the founder
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design. Re-docketed 2026-11-08: founder_reserved pattern extended to catch "pending the founder's decision"; record corrected from safe-default to founder-docket.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Harvest row 'Abu Bakr Al-Zabi' is an unresolved fragment: no contemporary dawah figure of this name exists in any source; the only near-match is the classical grammarian al-Zubaydi (d. 989). Should the row be logged unidentifiable, or does it resolve to the grammarian or another documented person? dec-20261108-169 · 2026-11-08
طے شدہ · rule · لین: modern-figures · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: log the item unidentifiable with every search recorded in the progress file — never guess a name, never invent an identity; a channel publishing under a name only gets a record under that publishing name
Answered by a standing repository rule (rule-unidentifiable): modern-figures lane no-skip rule + logged-unidentifiable precedent (Umat Muslim, the-muslim-mum).
Harvest row 'Karim Ismail' cannot be resolved to a single identity: the only large figure of the name is Kareem Esmail, a mental-health/life coach with no dawah channel. Log-and-hold per the ambiguous-name rule, or keep searching? dec-20261108-170 · 2026-11-08
طے شدہ · rule · لین: modern-figures · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: where there is doubt the classification is unknown/unclassified — doubt is not establishment; a name-basis or single weak attestation never settles a sensitive field; record the contested readings preserved in the note
Answered by a standing repository rule (rule-doubt-unknown): founder doubt rule, 24 Oct 2026 (docs/agents/lanes/modern-figures.md §3; revert commit 'founder re-review' 1 Nov 2026: 'the doubt rule is absolute').
Shirin Ebadi's record: EN Wikipedia quotes her own 'my family and I are Shiites' and the AR infobox gives مسلمة شيعية — two sources name Shia but no branch. The person-faith enum has no generic 'shia' value (only shia-twelver/ismaili/zaydi) and a branch was not inferred from Iranian origin. Add a generic 'shia' enum value, or keep such cases 'unknown' until a branch is explicitly cited? dec-20261108-171 · 2026-11-08
بانی کے پاس · policy-gate · لین: modern-figures · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
Rayta bint al-Harith's death place conflicts across the sira sources: one passage places her death in Abyssinia, another on the road — the record currently carries both as a debated reading. Is a debated dual-reading sufficient, or does this need a perspectives block? dec-20261108-172 · 2026-11-08
طے شدہ · rule · لین: hisham · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: contested topics use the perspectives block: Sunni position(s) first, then the academic view, then any other view — each cited and labelled, no verdict in the museum's voice
Answered by a standing repository rule (rule-perspectives): AGENTS.md (voice and scope).
Abdul Sancar's record has no Wikidata QID — nine indirect routes failed and the record went fail-closed per the wave-1 precedent. Should a single dedicated live-browser QID lookup pass be queued for Sancar (and Abdus Salam, Azim Premji), or is fail-closed permanent until a source surfaces one? dec-20261108-173 · 2026-11-08
طے شدہ · rule · لین: modern-figures · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: fix it in its own small commit: reproduce first, root-cause it, run validate + tests (+ build for site changes), append a CANARY entry — never hide a failure behind a guard or a widened timeout
Answered by a standing repository rule (rule-bug-fix): docs/agents/workflow.md + docs/ai-change-gates.md §5.
The Azhar academy book gives the Battle of Ayn Jalut as Friday 25 Ramadan 658 AH while the existing event entity gives Friday 15 Ramadan 658 — which Hijri date should the record carry, and how should the conflict be represented? dec-20261108-174 · 2026-11-08
طے شدہ · override · لین: azhar-academy · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Conflicting dates are preserved, not resolved silently: the entity keeps its recorded date and the book's variant is carried as an attributed alternative reading — never pick one date and bury the other.
dec-164 was generated minutes before the rule-merge pattern gained word boundaries: bare 'pr' matched inside 'represented' and 'conflict' matched the date-conflict phrasing, so a git-merge rule answered a Hijri-date question — the same false-positive class as dec-084/085/090. Pattern fixed; this override restores the correct doctrine (rule-conflict-preserve).
Abdul Sancar's record has no Wikidata QID — nine indirect routes failed and the record went fail-closed per the wave-1 precedent. Should a single dedicated live-browser QID lookup pass be queued for Sancar (and Abdus Salam, Azim Premji), or is fail-closed permanent until a source surfaces one? dec-20261108-175 · 2026-11-08
طے شدہ · override · لین: modern-figures · اعتماد: 1 · الٹنے کے قابل
فیصلہ: External IDs fail closed: never invent or guess a QID — the record stays without one until a source surfaces it. A single dedicated lookup pass may be queued as a process choice, but absence of an ID is never filled by inference.
dec-173 was generated minutes before the rule-bug-fix pattern gained word boundaries and a fail-closed guard: 'fail' inside 'fail-closed' plus 'API' inside an option matched the bug-fix workflow rule, so a broken-build doctrine answered an identifier-policy question. Pattern fixed; the engine now answers with rule-external-id-failclosed. This override restores that doctrine.
The authenticity validator requires named_edition (publisher_any + editor_any) for every accept decision. For pre-print canon books whose authenticity rests on manuscript tradition or institutional scans (no print publisher exists or is relevant), should accept still require a named modern edition, or should institutional provenance / a manuscript record satisfy the basis? dec-20261108-176 · 2026-11-08
طے شدہ · safe-default · لین: coverage-programme · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
EI2 queue row AMIR KHAN (p. 466-467) contains a headword inside it — AMIR KHUSRAW — with no queue row of its own. Should a new record be created for Amir Khusraw from that embedded headword, or is it covered by an existing record? dec-20261108-177 · 2026-11-08
طے شدہ · override · لین: ei2 · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Reuse before create: check whether the embedded headword is already covered by an existing entity; a duplicate becomes an alias of the richer record, and only a genuinely new subject gets a new record.
dec-163 cited rule-create-entity, but the match came from option text, not the question — after audit fix H5 (rules match question-only, like precedents) the engine returns safe-default. This override restores the reuse-first doctrine the record meant to apply; whether Amir Khusraw is already covered remains the lane's factual check.
The Azhar academy book's chapter on Ibtida' in religion overlaps the existing concept/bida entity — should the chapter's content extend concept/bida in place, or warrant a separate entity? dec-20261108-178 · 2026-11-08
طے شدہ · override · لین: azhar-academy · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Extend the existing concept/bida entity rather than mint a duplicate — overlapping subjects merge into the richer record with the new material cited there.
dec-167 cited rule-create-entity via option text only; post-H5 the engine returns safe-default. Override restores the reuse/extend doctrine for the ibtida-bida overlap.
A scholarly source claims consensus on a contested point without citing who agrees — soften to attributed claim or reject the source? dec-20261108-179 · 2026-11-08
طے شدہ · override · لین: admission · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Soften unsourced 'consensus' claims: if a consensus claim carries no citation it is not presented as consensus — attribute it, qualify it, or drop it.
dec-104's stored decision was an exclusions-doctrine misfire (same class as dec-129) yet the founder's label approved the intent — pairing question to wrong-rule decision teaches bad calibration. This override enacts the founder's actual ruling: soften unsourced consensus claims.
docs/coverage.md is stale after my entity batch — regenerate it with the coverage script or hand-edit the counts? dec-20261108-180 · 2026-11-08
طے شدہ · override · لین: docs · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Regenerate docs/coverage.md programmatically rather than hand-editing — generated docs are rebuilt by the pipeline, never edited by hand.
The founder verdict on dec-20261108-114 was approved — an affirmative ruling on a safe-default hold. Enacted as override for symmetry with the rejected-verdict overrides (audit L4): approved rulings become citable authority too.
My change edits a file under site/public/shell/ which rewrites all ~164k pages — label the PR rewrites-all-pages and batch it, or find a way to avoid the rewrite? dec-20261108-181 · 2026-11-08
طے شدہ · override · لین: deploy · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Shell-path churn is labelled: changes to files baked into every page carry the rewrites-all-pages label so the founder can batch deploys.
The founder verdict on dec-20261108-122 was approved — an affirmative ruling on a safe-default hold. Enacted as override for symmetry with the rejected-verdict overrides (audit L4): approved rulings become citable authority too.
Harvest row 'Karim Ismail' cannot be resolved to a single identity: the only large figure of the name is Kareem Esmail, a mental-health/life coach with no dawah channel. Log-and-hold per the ambiguous-name rule, or keep searching? dec-20261108-182 · 2026-11-08
طے شدہ · override · لین: modern-figures · اعتماد: 1 · الٹنے کے قابل
فیصلہ: An unresolved harvest/identity row is logged and held for a dedicated lookup pass — the engine does not settle identity by guessing. (Correct rule: rule-unidentifiable; rule-doubt-unknown fired in error on priority.)
Founder approved the outcome (yes) but the cited rule was wrong — doubt-unknown answered an identity question. Corrected per Muse/Claude audit; pattern-priority note for the engine.
The authenticity validator requires named_edition (publisher_any + editor_any) for every accept decision. For pre-print canon books whose authenticity rests on manuscript tradition or institutional scans (no print publisher exists or is relevant), should accept still require a named modern edition, or should institutional provenance / a manuscript record satisfy the basis? dec-20261108-183 · 2026-11-08
طے شدہ · override · لین: coverage-programme · اعتماد: 1 · الٹنے کے قابل
فیصلہ: accept requires authenticity_basis (named edition OR institutional provenance OR manuscript/primary record — policy §6's disjunctive bases). named_edition is NOT a condition of the decision; it stays mandatory at the harvest gate (authenticity_gate.py) where its title-matching function lives.
Founder ruling 2026-11-08, option (b). A manuscript basis is more primary than a print edition; forcing modern publisher metadata on pre-print works was an architectural error (validator exceeded the policy it implements). check-authenticity.ts updated accordingly.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Does one independent T2–T4 school source suffice for an authenticity accept, or are two required (policy §8)? dec-20261108-184 · 2026-11-08
طے شدہ · override · لین: delegation · اعتماد: 1 · الٹنے کے قابل
فیصلہ: One independent T2–T4 source suffices (preferably encyclopaedic/institutional). This records the founder ruling of 31 Oct 2026 enacted in commit 1c829f6da5 (PR #2816), confirmed by the founder on 2026-11-08 — policy §8 amended to match.
Founder ruled 1a on 2026-11-08 after Claude located the traceable commit. §8's two-source text was stale relative to the enacted register rule.
My PR is green on validate and test — should I merge it myself or wait for a reviewer? dec-20261108-185 · 2026-11-08
طے شدہ · override · لین: workflow · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Agents merge their own PRs automatically once gates pass — but only after fetching origin/main, reviewing what landed, and three-way merging so nothing is overwritten. Deployment stays founder-manual. (Narrows dec-160: the no-self-merge reading was broader than the founder's wave-2 words, which concerned decision-ledger records.)
Founder ruling Q8, 2026-11-08: 'allow all agents to automatically merge but they all have to first get latest from main and understand if any changes and check where to fit and merge their changes.' AGENTS.md canary-workflow and PR-rule text updated to match.
Who labels decision records going forward, and what is the authority ceiling for agent-ruled precedents? dec-20261108-186 · 2026-11-08
طے شدہ · override · لین: delegation · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The founder delegates ALL decision-record labeling to the Devin agent effective 2026-11-08: the agent applies founder-equivalent verdicts (approved/held/rejected) based on learned doctrine from prior founder rulings, without per-item founder review. Agent-ruled-delegated precedents retain FULL deciding authority (no cap reduction) — full autonomy is the explicit goal. TTL/supersession machinery is added to precedents.yaml as a kill-switch for dead rulings, not an authority reduction. Founder batch-approval for canon batches (dec-133) likewise trends toward full agent autonomy.
Founder rulings Q6/Q9/dec-133-note, 2026-11-08: 'i will not be labeling from now on, you do the labeling'; 'i want to give full authority to agents and not reduced at all, this is the goal here full autonomity'; 'my goal is that i don't even need to approve batches even.' Hard gates (deploy/spend/outreach/delete/content-safety/Qur'an/governance surfaces) remain founder-docket regardless — autonomy does not extend to non-delegable gates.
How are the remaining ~286 undecided canon books decided under the new delegation? dec-20261108-187 · 2026-11-08
طے شدہ · override · لین: delegation · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: All 292 undecided canon books decided by agent in one sweep (2026-11-08): 234 accept (school sunni verified via person-graph traditions or attested-author research), 61 label (non-Sunni: 1 shia, 5 other/contested; unverified: ~55). Evidence basis per dec-183 (manuscript/institutional provenance + harvested carriers); school evidence cites internal person records or attested-author notes. Uncertain cases default to label/unverified — doubt stays labelled, never silently accepted.
First full exercise of dec-186 delegation. Method: strict person-graph match (exact normalized Arabic names only — last-token matching rejected after it produced wrong-person hits) + attested-author map for famous scholars + book-level overrides for anonymous compilations. Every register entry decided_by: devin-swe2 (agent, founder delegation dec-20261108-186). Labels are reversible — upgrade path exists when second sources are gathered (phase-13 wikidata-second-source).
What is the reference set that rules and the rule engine must conform to? dec-20261108-188 · 2026-11-08
طے شدہ · override · لین: delegation · اعتماد: 1 · الٹنے کے قابل نہیں
فیصلہ: GOLDEN SET RULE (founder, 2026-11-08): everything currently in islam.museum — all data, prose, events, persons, books — is the golden reference. Rules and the decision engine are calibrated TO it, never the reverse. If a rule would rule out an existing entity, the rule/engine is wrong and must be adjusted to accommodate the golden set. Consequences: (1) existing published content is never retroactively invalidated by new rules — new rules apply forward unless the founder says otherwise; (2) decision-engine rules that conflict with attested existing content get fixed, not the content; (3) canon labels/decisions must be consistent with how the museum already treats these authors and works.
Founder ruling, verbatim intent: 'everything that is right now in islam museum... is your golden set... if your rules rule out an existing book, prose, phrase, event, anything then the rule and the rule engine is wrong and needs to be adjusted.' This supersedes any interpretation where new doctrine silently invalidates existing content — including the draft-tier AI rule (applies to new AI output only, existing published content stands).
Should alfehrest.org (AlFehrest — interactive timeline/map of Islamic civilisation 700-1500 CE: states, rulers, scholars, events) be onboarded as a citable source family, and if not, at what registration tier? Assessment evidence: volunteer project, no institution/editorial board, self-disclaims academic-reference status; stable numeric IDs for states/scholars; Arabic-primary; ~200 scholar bios and polity succession lists derived largely from Ibn Khaldun, al-Dhahabi Siyar, Mu'nis atlas, Ibn Kathir, plus Wikipedia/Hukam.net; GitHub JSON datasets last touched 2018 under informal CC-Share-Alike wording; coverage stops at 1500 CE. Name collision: unrelated to FIHRIST (fihrist.org.uk), the Oxford/Cambridge manuscript union catalogue. dec-20261108-189 · 2026-11-08
طے شدہ · rule · لین: library · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should alfehrest.org (AlFehrest — interactive timeline/map of Islamic civilisation 700-1500 CE: states, rulers, scholars, events) be onboarded as a citable source family, and if not, at what registration tier? Assessment evidence: volunteer project, no institution/editorial board, self-disclaims academic-reference status; stable numeric IDs for states/scholars; Arabic-primary; ~200 scholar bios and polity succession lists derived largely from Ibn Khaldun, al-Dhahabi Siyar, Mu'nis atlas, Ibn Kathir, plus Wikipedia/Hukam.net; GitHub JSON datasets last touched 2018 under informal CC-Share-Alike wording; coverage stops at 1500 CE. Name collision: unrelated to FIHRIST (fihrist.org.uk), the Oxford/Cambridge manuscript union catalogue. dec-20261108-190 · 2026-11-08
طے شدہ · override · لین: library · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 3 pointer/orientation only. Register AlFehrest in docs/agents/source-families.md alongside Encyclopaedia Iranica: use it for discovery and diffing (polity successions, snapshot completeness, scholar whereabouts 700-1500 CE); every fact re-cited to the underlying source each page names. Never a {cite:} target, no data/source entity, no harvest-registry entry. GitHub state JSONs usable as facts-only comparison data — never republished into CC0. Record the FIHRIST (fihrist.org.uk) name collision in the entry. Separately assess IDEO Diamond / al-Kindi (ideo.diamondrda.org, Dominican Institute for Oriental Studies) as a works/authors authority-file candidate for the 1,153-work prose backlog.
dec-189 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment evidence ('Arabic-primary', 'Arabic is the primary'). The rule fired on descriptive metadata, not on a language/localisation question — same substring false-positive class as the kid/wikidata bug. The correct ruling applies the eight onboarding criteria: authority fails (volunteer project, self-disclaimed non-reference), so Tier 3 pointer is the ceiling; the assessment's own recommendation is adopted verbatim in scope.
Should IDEO Diamond / AlKindi (opac.diamondrda.org — the IFLA-LRM/RDA catalogue run by IDEO Cairo and stewarded by LRM-CS: >230,000 works, >116,000 agents, typed work-to-work relations) be onboarded as a source family for the works backlog, and at what tier? Assessment evidence: institutional scholarly authority (IDEO est. 1953, Vatican Library and Arab League manuscript institute among 11 partners); stable numeric work/agent IDs already used by Wikidata properties P8795 and P11979; no published data licence — open-data ethos only; no public API, SPARQL endpoint or bulk download found; coverage concentrated in the classical Islamicate heritage the museum mines. Proposed family id source/ideo-diamond, citation anchor work/<id> or agent/<id>. dec-20261108-191 · 2026-11-08
بانی کے پاس · policy-gate · لین: library · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: licence changes are submission-impacting
Should tafsir.app (al-Bahith al-Qur'ani, Nuqayah / Jam'iyyat Ayat) be harvested into the local workspace for mining — ~80 text books (41 tafsir, 30 ulum al-quran, 9 ma'ajim) via its own get.php JSON endpoint carrying printed-page markers — under an unlicensed/facts-only posture? Recon verified 2026-11-08: no licence or terms page exists; texts reproduce modern critical editions (e.g. Shakir's Tabari tahqiq apparatus); endpoint is the site's own public route used by its web app; robots/terms check pending in Prompt 0. Harvested texts would live ONLY in the harvest workspace, never enter git or R2, and citations would name the underlying book/edition — {cite:tafsir-app} registration deferred to a separate founder decision. dec-20261108-192 · 2026-11-08
بانی کے پاس · founder-reserved · لین: library · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
Should the British Museum collection database (britishmuseum.org/collection — object records with museum numbers, dates, materials, provenance for the Islamic collections) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (national museum); citability good (museum object numbers); terms: collection data published CC BY-NC-SA 4.0 per the family catalogue's earlier check — non-commercial/share-alike clause means facts-only posture permanently; structure: collection pages answer 403 Cloudflare challenge to non-browser clients on 2026-11-08 (recorded, not bypassed); Wayback snapshots exist; an official data release/API status needs a browser check; coverage fit good (BM Islamic collection vs 542 object records); corroboration high; Arabic-first moderate; longevity strong. dec-20261108-194 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the Khalili Collections online database (khalilicollections.org — the Nasser D. Khalili family's published Islamic-art collections: metalwork, ceramics, manuscripts, arms, textiles, with per-object pages under /collections/islamic-art/<id>-<slug>/) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority good (named specialist collection, published catalogues exist in print); citability good (stable object page ids); terms: no licence or terms page found — default facts-only; structure: WordPress HTML pages plus XML sitemap, no API; robots file allows all with sitemap; coverage fit good (major Islamic decorative-arts collection vs 542 object records); corroboration moderate-high; Arabic-first low-moderate; longevity moderate (private foundation site). dec-20261108-195 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the Qatar Museums unified collections platform (collections.qm.org.qa — hosts the Museum of Islamic Art Doha object database with faceted search and per-object pages at /objects/<slug>, plus other QM collections) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (Qatar Museums national institution); citability good (stable object slugs, e.g. /objects/lusterware-bowl-mia201373); terms: terms page at qm.org.qa/en/terms-and-conditions — posture to record as facts-only unless it clearly permits reuse; structure: server-rendered faceted search with pagination, JSON state embedded in pages, no documented public API found yet; coverage fit good (MIA Doha is a major Islamic collection vs 542 object records); corroboration high (provenance, dated works); Arabic-first moderate; longevity strong. dec-20261108-196 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the Thesaurus d'Epigraphie Islamique (now at epigraphie-islamique.uliege.be/Thesaurus/, University of Liege hosting, Max van Berchem Foundation sponsorship, directed by named scholars Kalus/Bauden/Soudan, successor to the printed RCEA) be onboarded as a citable source family for dating buildings, objects and events, and at what tier? Assessment 2026-11-08: authority strong; citability excellent (stable fiche numbers, e.g. EpigraphyDisplay.aspx?id=11979, GET pages server-rendered); terms: free of charge online since Oct 2011, no licence page found — default facts-only; structure: ASP.NET postback search (scriptable) plus GET fiche pages carrying Hijri era, Christian era, place, current repository, language, nature, support type, material, last-update date; 58,289 fiches and 49,884 photos covering the Muslim world to AH 1000; coverage fit high (inscription dating is the corroboration the corpus lacks); Arabic-first strong; longevity strong (university host, live daily updates). Old domain epigraphie-islamique.org is a parked courtesy page — the uliege.be base is canonical. dec-20261108-197 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the UNESCO World Heritage Centre property records and Documents pages (whc.unesco.org — per-property dossiers: nomination files, ICOMOS evaluations, criteria, inscription dates) be onboarded as a citable source family for listed Islamic-world properties, and at what tier? Assessment 2026-11-08: authority strong (UNESCO); citability excellent (property numbers /en/list/<n> and document ids /document/<n>); terms: public documents published for consultation; access: all HTML pages answer 403 Cloudflare challenge to non-browser clients on 2026-11-08 (robots file itself served 200); Wayback Machine holds snapshots; coverage fit moderate (only WHC-listed properties, a subset of the building backlog); corroboration high (expert-written dated dossiers); Arabic-first moderate; longevity strong. dec-20261108-198 · 2026-11-08
بانی کے پاس · policy-gate · لین: web-sources · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: external communications, commissions and commitments are founder-only — negation tricks must not disarm this gate
Should Archnet (archnet.org — the Aga Khan Trust for Culture / MIT Aga Khan Documentation Center open resource on the architecture and built environment of Muslim societies, ~9,000 site records, ~1,000 authority records, 120k+ media items) be onboarded as a citable source family for the building records, and at what tier? Assessment 2026-11-08: authority strong (AKTC+MIT joint project since 2002, peer-reviewed claims on its about page); citability good (stable numeric site/authority/media ids, e.g. /sites/1000); terms: its Terms of Use page states material is 'available for instructional, nonprofit use, and personal use on the terms of a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license' and its help page says site descriptions and metadata may be used by nonprofit or educational organisations with attribution — non-commercial clause means facts-only posture permanently; structure: live site answers 429 to non-browser clients (Vercel checkpoint), archived robots file has no disallow rules, Wayback Machine holds server-rendered record pages, no public data API found; coverage fit excellent (5,776 building records); corroboration high (construction dates, architects, site types); Arabic-first moderate; longevity strong. dec-20261108-199 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the Victoria and Albert Museum Collections API (api.vam.ac.uk/v2 — public JSON search returning systemNumber, accessionNumber, dates, maker, place, object type and IIIF image identifiers; collections.vam.ac.uk object pages) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (national museum); citability excellent (systemNumber stable id, IIIF manifests); terms: V&A websites terms state V&A-owned content may be used free of charge for non-commercial purposes — facts-only posture for data/prose, images never ingested; structure: keyless JSON API v2 answered live (objects/search?q=minbar returned records); robots file names AI crawlers only with no wildcard group, harvest UA not listed; coverage fit good (V&A Islamic collection vs 542 object records); corroboration high (maker/date/material/provenance); Arabic-first moderate; longevity strong. dec-20261108-200 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the British Museum collection database (britishmuseum.org/collection — object records with museum numbers, dates, materials, provenance for the Islamic collections) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (national museum); citability good (museum object numbers); terms: collection data published CC BY-NC-SA 4.0 per the family catalogue's earlier check — non-commercial/share-alike clause means facts-only posture permanently; structure: collection pages answer 403 Cloudflare challenge to non-browser clients on 2026-11-08 (recorded, not bypassed); Wayback snapshots exist; an official data release/API status needs a browser check; coverage fit good (BM Islamic collection vs 542 object records); corroboration high; Arabic-first moderate; longevity strong. dec-20261108-201 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:british-museum <object-id>}
dec-194 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should the Khalili Collections online database (khalilicollections.org — the Nasser D. Khalili family's published Islamic-art collections: metalwork, ceramics, manuscripts, arms, textiles, with per-object pages under /collections/islamic-art/<id>-<slug>/) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority good (named specialist collection, published catalogues exist in print); citability good (stable object page ids); terms: no licence or terms page found — default facts-only; structure: WordPress HTML pages plus XML sitemap, no API; robots file allows all with sitemap; coverage fit good (major Islamic decorative-arts collection vs 542 object records); corroboration moderate-high; Arabic-first low-moderate; longevity moderate (private foundation site). dec-20261108-202 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:khalili-collections <object-id>}
dec-195 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should the Qatar Museums unified collections platform (collections.qm.org.qa — hosts the Museum of Islamic Art Doha object database with faceted search and per-object pages at /objects/<slug>, plus other QM collections) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (Qatar Museums national institution); citability good (stable object slugs, e.g. /objects/lusterware-bowl-mia201373); terms: terms page at qm.org.qa/en/terms-and-conditions — posture to record as facts-only unless it clearly permits reuse; structure: server-rendered faceted search with pagination, JSON state embedded in pages, no documented public API found yet; coverage fit good (MIA Doha is a major Islamic collection vs 542 object records); corroboration high (provenance, dated works); Arabic-first moderate; longevity strong. dec-20261108-203 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:qatar-museums-collections <object-slug>}
dec-196 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should the Thesaurus d'Epigraphie Islamique (now at epigraphie-islamique.uliege.be/Thesaurus/, University of Liege hosting, Max van Berchem Foundation sponsorship, directed by named scholars Kalus/Bauden/Soudan, successor to the printed RCEA) be onboarded as a citable source family for dating buildings, objects and events, and at what tier? Assessment 2026-11-08: authority strong; citability excellent (stable fiche numbers, e.g. EpigraphyDisplay.aspx?id=11979, GET pages server-rendered); terms: free of charge online since Oct 2011, no licence page found — default facts-only; structure: ASP.NET postback search (scriptable) plus GET fiche pages carrying Hijri era, Christian era, place, current repository, language, nature, support type, material, last-update date; 58,289 fiches and 49,884 photos covering the Muslim world to AH 1000; coverage fit high (inscription dating is the corroboration the corpus lacks); Arabic-first strong; longevity strong (university host, live daily updates). Old domain epigraphie-islamique.org is a parked courtesy page — the uliege.be base is canonical. dec-20261108-204 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:tei-epigraphie <fiche-id>}
dec-197 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should Archnet (archnet.org — the Aga Khan Trust for Culture / MIT Aga Khan Documentation Center open resource on the architecture and built environment of Muslim societies, ~9,000 site records, ~1,000 authority records, 120k+ media items) be onboarded as a citable source family for the building records, and at what tier? Assessment 2026-11-08: authority strong (AKTC+MIT joint project since 2002, peer-reviewed claims on its about page); citability good (stable numeric site/authority/media ids, e.g. /sites/1000); terms: its Terms of Use page states material is 'available for instructional, nonprofit use, and personal use on the terms of a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license' and its help page says site descriptions and metadata may be used by nonprofit or educational organisations with attribution — non-commercial clause means facts-only posture permanently; structure: live site answers 429 to non-browser clients (Vercel checkpoint), archived robots file has no disallow rules, Wayback Machine holds server-rendered record pages, no public data API found; coverage fit excellent (5,776 building records); corroboration high (construction dates, architects, site types); Arabic-first moderate; longevity strong. dec-20261108-205 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:archnet <site|authority|media-id>} — Wayback/browser read channel while live site rate-limits non-browser clients
dec-199 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should the Victoria and Albert Museum Collections API (api.vam.ac.uk/v2 — public JSON search returning systemNumber, accessionNumber, dates, maker, place, object type and IIIF image identifiers; collections.vam.ac.uk object pages) be onboarded as a citable source family for the object backlog, and at what tier? Assessment 2026-11-08: authority strong (national museum); citability excellent (systemNumber stable id, IIIF manifests); terms: V&A websites terms state V&A-owned content may be used free of charge for non-commercial purposes — facts-only posture for data/prose, images never ingested; structure: keyless JSON API v2 answered live (objects/search?q=minbar returned records); robots file names AI crawlers only with no wildcard group, harvest UA not listed; coverage fit good (V&A Islamic collection vs 542 object records); corroboration high (maker/date/material/provenance); Arabic-first moderate; longevity strong. dec-20261108-206 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:vam-collections <systemNumber>}; keyless JSON API is the harvest channel
dec-200 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first value', 'Arabic-first moderate'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 (kid/wikidata and Arabic-in-evidence). The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms posture, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; the assessment's recommended Tier 1 facts-only posture is adopted (non-commercial or unclear terms => facts-only permanently).
Should the Arabic Papyrology Database (now at apd.gwi.uni-muenchen.de/apd/, LMU Munich, ISAP patronage, Trismegistos partner) be onboarded as a citable source family for documentary evidence of early Islamic Egypt, and at what tier? dec-20261108-207 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should HMML (Hill Museum & Manuscript Library, hmml.org — catalogue of 60,000+ manuscript records with shelfmarks, dates and provenance notes) be onboarded as a citable source family for manuscript dating and provenance, and at what tier? dec-20261108-208 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should Manar al-Athar (manar.arch.ox.ac.uk — University of Oxford digital archive of ~83,000 labelled photographs of Islamic-period archaeological sites, buildings and art, bilingual labels) be onboarded as a citable source family for building and site records, and at what tier? dec-20261108-209 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should the Museum With No Frontiers 'Discover Islamic Art' database (islamicart.museumwnf.org) and the Qantara Mediterranean heritage catalogue (qantara-med.org — object and monument records across the Mediterranean, overlapping the Sicily material) be onboarded as citable source families, and at what tier? dec-20261108-210 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should the Princeton Geniza Project (geniza.princeton.edu — 4,350+ transcribed Cairo Geniza documents with published metadata) be onboarded as a citable source family for Mediterranean trade, Fustat daily life and the Sicily route, and at what tier? dec-20261108-211 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should the Qatar Digital Library (qdl.qa — manuscripts, documents and oral histories from the Islamic world, including dated copies with colophons) be onboarded as a citable source family for work dating and provenance, and at what tier? dec-20261108-212 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should Zeno.ru (Oriental Coins Database — coin records with stable ids cross-referenced to Album/Tye/Mitchiner catalogues) be onboarded as a citable source family for polity and ruler coinage, and at what tier? dec-20261108-213 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Should the Arabic Papyrology Database (now at apd.gwi.uni-muenchen.de/apd/, LMU Munich, ISAP patronage, Trismegistos partner) be onboarded as a citable source family for documentary evidence of early Islamic Egypt, and at what tier? dec-20261108-214 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:apd <document-id>}; non-commercial project — record text never republished
dec-207 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should HMML (Hill Museum & Manuscript Library, hmml.org — catalogue of 60,000+ manuscript records with shelfmarks, dates and provenance notes) be onboarded as a citable source family for manuscript dating and provenance, and at what tier? dec-20261108-215 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 2 citable family, facts-only: cite {cite:hmml <record-id>} — public catalogue metadata only; no login-gated content
dec-208 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should Manar al-Athar (manar.arch.ox.ac.uk — University of Oxford digital archive of ~83,000 labelled photographs of Islamic-period archaeological sites, buildings and art, bilingual labels) be onboarded as a citable source family for building and site records, and at what tier? dec-20261108-216 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:manar-al-athar <record-id>} — captions and identifications only; NC licence means images never ingested
dec-209 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should the Museum With No Frontiers 'Discover Islamic Art' database (islamicart.museumwnf.org) and the Qantara Mediterranean heritage catalogue (qantara-med.org — object and monument records across the Mediterranean, overlapping the Sicily material) be onboarded as citable source families, and at what tier? dec-20261108-217 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Both Tier 2 citable families, facts-only: cite {cite:mwnf-discover-islamic-art <id>} and {cite:qantara-med <id>}; terms unconfirmed — record text never republished
dec-210 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should the Princeton Geniza Project (geniza.princeton.edu — 4,350+ transcribed Cairo Geniza documents with published metadata) be onboarded as a citable source family for Mediterranean trade, Fustat daily life and the Sicily route, and at what tier? dec-20261108-218 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:princeton-geniza <document-id>}; Zenodo metadata is the machine channel
dec-211 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should the Qatar Digital Library (qdl.qa — manuscripts, documents and oral histories from the Islamic world, including dated copies with colophons) be onboarded as a citable source family for work dating and provenance, and at what tier? dec-20261108-219 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:qdl <document-id>} — Wayback/browser read channel while live site challenges non-browser clients
dec-212 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should Zeno.ru (Oriental Coins Database — coin records with stable ids cross-referenced to Album/Tye/Mitchiner catalogues) be onboarded as a citable source family for polity and ruler coinage, and at what tier? dec-20261108-220 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 3 pointer only — catalogue ids locate the published reference; robots.txt disallows all automated access and the database is derivative of published coin catalogues
dec-213 answered a source-onboarding question with rule-source-families, whose pattern 'families? of sources?'/'source famil' matched the phrase 'source family' describing the candidate itself. The rule governs how attestations are COUNTED (one family = one attestation), not whether a family may be registered — same substring false-positive class as dec-189->190 and dec-194..200->201..206. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md; the assessment's recommended tier and posture is adopted (non-commercial or unclear terms => facts-only permanently; robots or login barriers cap the channel, not necessarily the tier).
Should the EAMENA database (Endangered Archaeology in the Middle East and North Africa — University of Oxford with Leicester and Durham, Getty/WMF ARCHES platform, database.eamena.org, ~383,000 heritage-place records) be onboarded as a citable source family for the building/place backlog, and at what tier? Assessment 2026-11-08: authority excellent (three universities + Getty Conservation Institute platform); citability good (Heritage Place record ids, search URLs, Zenodo dataset community); terms: site states 'The EAMENA Database is licensed under a Creative Commons Attribution 4.0 International License' — open with attribution, BUT access is tiered: Public Access withholds coordinates, zoom and condition data; Research Access needs registration; Contributor Access for edits; structure excellent (ARCHES 7 API, GeoJSON export endpoint, citation plugin to Zenodo); coverage fit exceptional (~383k MENA archaeological site records vs 5,460 prose-missing buildings); corroboration high (site condition assessments, period data); Arabic-first moderate (bilingual EN/AR interface); longevity strong (university + Zenodo). dec-20261108-221 · 2026-11-08
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should the EAMENA database (Endangered Archaeology in the Middle East and North Africa — University of Oxford with Leicester and Durham, Getty/WMF ARCHES platform, database.eamena.org, ~383,000 heritage-place records) be onboarded as a citable source family for the building/place backlog, and at what tier? Assessment 2026-11-08: authority excellent (three universities + Getty Conservation Institute platform); citability good (Heritage Place record ids, search URLs, Zenodo dataset community); terms: site states 'The EAMENA Database is licensed under a Creative Commons Attribution 4.0 International License' — open with attribution, BUT access is tiered: Public Access withholds coordinates, zoom and condition data; Research Access needs registration; Contributor Access for edits; structure excellent (ARCHES 7 API, GeoJSON export endpoint, citation plugin to Zenodo); coverage fit exceptional (~383k MENA archaeological site records vs 5,460 prose-missing buildings); corroboration high (site condition assessments, period data); Arabic-first moderate (bilingual EN/AR interface); longevity strong (university + Zenodo). dec-20261108-222 · 2026-11-08
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family under CC BY 4.0 with attribution. Register source/eamena: registry entry, data/source entity with reliability_note recording the licence statement, the three access tiers and the public-tier limits (no coordinates, zoom or condition data without Research registration — so no coordinate claims sourced from the public tier), attribution row in docs/attribution.md, Tier-1 table row in docs/agents/source-families.md. Harvest channel: ARCHES search + GeoJSON export (database.eamena.org/api/search/export_results) and Zenodo dataset community; cite {cite:eamena <heritage-place-id>} — the ARCHES record id, stable and reader-resolvable. Pilot on 10 building/place records. The CC BY licence permits reuse with attribution — the museum's citation is the attribution; record text still enters data/ in the museum's own words per the standing sources policy.
dec-221 answered a source-onboarding question with the Arabic-first content rule because the assessment evidence mentioned 'Arabic-first moderate' and the bilingual interface. The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-190 (AlFehrest) and the kid/wikidata fix. The correct ruling applies the eight onboarding criteria: authority, citability, stated terms (CC BY 4.0), structure (ARCHES API + GeoJSON export), coverage fit (~383k MENA site records against the 5,460-building prose gap), corroboration, bilingual interface and institutional longevity all pass; the access tiers are recorded as limits, not blockers.
Should Nomisma.org (the American Numismatic Society / Oxford-linked Linked Open Data ontology for numismatic concepts, CC-BY 3.0, with SPARQL endpoint and documented APIs) be onboarded as a citable source family for Islamic coinage, mints, issuing authorities and regnal dates, and at what tier? Assessment 9 Nov 2026: authority excellent (ANS + Oxford + INC patronage since 2010, named scientific committee); citability excellent (stable nomisma.org/id/ URIs per concept); terms CC-BY 3.0 stated on site with per-dataset licence checks needed; structure excellent (SPARQL + typed APIs + RDFa); coverage fit excellent and urgent (kind: coin objects ~2 in the graph while Islamic numismatics spans ~1,300 years of mints/reigns — coins corroborate place/person/polity across the whole timeline); corroboration excellent (coins are primary evidence for mint towns, regnal dates and titulature); Arabic-first moderate (legends Arabic, labels primarily en/de); longevity excellent (ANS institutional hosting since 2010). dec-20261109-223 · 2026-11-09
طے شدہ · rule · لین: web-sources · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Arabic first (label.ar + summary.ar required); English and Urdu in the same commit
Answered by a standing repository rule (rule-language-ar-first): docs/language.md + AGENTS.md.
Should Nomisma.org (the American Numismatic Society / Oxford-linked Linked Open Data ontology for numismatic concepts, CC-BY 3.0, with SPARQL endpoint and documented APIs) be onboarded as a citable source family for Islamic coinage, mints, issuing authorities and regnal dates, and at what tier? Assessment 9 Nov 2026: authority excellent (ANS + Oxford + INC patronage since 2010, named scientific committee); citability excellent (stable nomisma.org/id/ URIs per concept); terms CC-BY 3.0 stated on site with per-dataset licence checks needed; structure excellent (SPARQL + typed APIs + RDFa); coverage fit excellent and urgent (kind: coin objects ~2 in the graph while Islamic numismatics spans ~1,300 years of mints/reigns — coins corroborate place/person/polity across the whole timeline); corroboration excellent (coins are primary evidence for mint towns, regnal dates and titulature); Arabic-first moderate (legends Arabic, labels primarily en/de); longevity excellent (ANS institutional hosting since 2010). dec-20261109-224 · 2026-11-09
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 2 citable family: cite {cite:nomisma <concept-id>} for numismatic facts (mint, issuer, type, standard, date); partner-dataset records re-checked per-record for licence posture
dec-223 answered a source-onboarding question with the Arabic-first content rule because the word 'Arabic' appeared in the assessment text ('Arabic-first moderate', 'legends Arabic'). The rule fired on descriptive metadata, not on a language/localisation question — the same substring false-positive class as dec-189->190 and dec-197->204. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: authority, citability, terms, structure, coverage fit, corroboration, Arabic coverage and longevity all scored; Nomisma's own CC-BY 3.0 licence permits citation with attribution ({cite:nomisma <concept-id>} is the attribution), while partner datasets ingested through Nomisma carry their own licence statements and stay facts-only until checked per-record — hence Tier 2, not Tier 1.
Should the ANS MANTIS collection database (American Numismatic Society's ~98k-object coin catalogue, ~60,672 records in the Islamic department, ODbL-licensed data with stable accession ids, NUDS/RDF/JSON-LD/Atom APIs, Nomisma-linked authorities) be onboarded as a citable source family for coin objects, and at what tier? Assessment 9 Nov 2026: authority excellent (ANS is the premier US numismatic institution; records carry agency, publication status and maintenance history); citability excellent (stable accession ids like 1917.215.2009, resolvable per-record in NUDS/XML, RDF/XML, Turtle, JSON-LD, KML); terms good — data ODbL, images Public Domain Mark / No Copyright-US; structure excellent (documented REST + Atom feeds + Solr/Lucene queries, per-record serializations); coverage fit excellent and urgent — the graph holds only 2 kind:coin objects while MANTIS carries tens of thousands of Islamic coins spanning the whole timeline, and records already link Nomisma authority/mint URIs matching the museum's external_ids.nomisma; corroboration excellent (primary physical evidence for issuers, mints, denominations, dates, weights); language coverage moderate (legends predominantly in Arabic script, labels English); longevity excellent (institutional database since ~2007, actively maintained). dec-20261109-225 · 2026-11-09
بانی کے پاس · policy-gate · لین: web-sources · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: external communications, commissions and commitments are founder-only — negation tricks must not disarm this gate
Should the ANS MANTIS collection database (American Numismatic Society's ~98k-object coin catalogue, ~60,672 records in the Islamic department, ODbL-licensed data with stable accession ids, NUDS/RDF/JSON-LD/Atom APIs, Nomisma-linked authorities) be onboarded as a citable source family for coin objects, and at what tier? Assessment 9 Nov 2026: authority excellent (ANS is the premier US numismatic institution; records carry agency, publication status and maintenance history); citability excellent (stable accession ids like 1917.215.2009, resolvable per-record in NUDS/XML, RDF/XML, Turtle, JSON-LD, KML); terms good — data ODbL, images Public Domain Mark / No Copyright-US; structure excellent (documented REST + Atom feeds + Solr/Lucene queries, per-record serializations); coverage fit excellent and urgent — the graph holds only 2 kind:coin objects while MANTIS carries tens of thousands of Islamic coins spanning the whole timeline, and records already link Nomisma authority/mint URIs matching the museum's external_ids.nomisma; corroboration excellent (primary physical evidence for issuers, mints, denominations, dates, weights); language coverage moderate (legends predominantly in Arabic script, labels English); longevity excellent (institutional database since ~2007, actively maintained). dec-20261109-226 · 2026-11-09
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 1 citable family, facts-only: cite {cite:ans-mantis <accession-id>} for object-level numismatic facts (denomination, mint, issuer, date, weight, fabric); catalogue prose and legend text never republished; bulk ingestion deferred pending ODbL share-alike review — pilot on a bounded set of coins of already-linked rulers first
dec-225 answered a source-onboarding question with the no-decision outreach gate because the substring 'nomination' appears inside 'denominations' — the coin denomination field named in the assessment and the proposed citation form. The gate fired on a lexical accident, not on any external-communication, commission or commitment content; this is the same substring false-positive class as dec-189->190 (kid/wikidata), dec-197->204 and dec-222->224 (Arabic-in-evidence), and the fourth live specimen for the rule-precision audit. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: the American Numismatic Society is the premier US numismatic institution with per-record maintenance history; accession ids are stable and resolvable in six serializations; the data licence is a real open licence (ODbL) with Public Domain Mark / No Copyright-US images; the API surface is documented REST plus Atom feeds; coverage is the missing coin object class itself (~60,672 Islamic-department records against 2 kind:coin objects today); records already carry Nomisma authority/mint URIs matching the museum's external_ids.nomisma. Facts-only is chosen over bulk ingestion because ODbL share-alike attaches to derivative databases — citation of facts with attribution is clean, wholesale structured ingest needs its own licence review.
Should the iDAI.gazetteer (the German Archaeological Institute's ancient-world place gazetteer, canonical place/<gazId> URIs, unauthenticated JSON REST endpoints, CC-BY 4.0 content licence, multilingual names including Arabic, provenance-tracked records with owl:sameAs links to Pleiades/geonames/GND) be onboarded as a citable source family for place records, and at what tier? Assessment 9 Nov 2026: authority excellent (the DAI is Germany's federal archaeology institute and the gazetteer is a Pelagios linked-data standard across the ancient-world research community); citability excellent (stable canonical URIs like https://gazetteer.dainst.org/place/2280221 for Damascus, gazId + doc/<id>.json full record); terms good — CC-BY 4.0 stated in the site footer; structure excellent — robots.txt 'User-agent: * / Allow: /', JSON endpoints search.json, doc/<id>.json, suggestions, children, geoSearch; coverage fit good — populated-place, archaeological-site and building-institution types with coordinates and polygons across the Islamic-period Mediterranean and Near East; corroboration good — per-record identifiers (zenon-systemnr, zenon-thesaurus, GND-ID) and owl:sameAs links to Pleiades/geonames, the same identifier families the museum's place records already carry as external_ids; language coverage good — names in about 25 languages including Arabic; longevity excellent — institutional DAI service, versioned releases, per-record provenance batches. dec-20261109-227 · 2026-11-09
بانی کے پاس · policy-gate · لین: web-sources · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
Should the iDAI.gazetteer (the German Archaeological Institute's ancient-world place gazetteer, canonical place/<gazId> URIs, unauthenticated JSON REST endpoints, CC-BY 4.0 content licence, multilingual names including Arabic, provenance-tracked records with owl:sameAs links to Pleiades/geonames/GND) be onboarded as a citable source family for place records, and at what tier? Assessment 9 Nov 2026: authority excellent (the DAI is Germany's federal archaeology institute and the gazetteer is a Pelagios linked-data standard across the ancient-world research community); citability excellent (stable canonical URIs like https://gazetteer.dainst.org/place/2280221 for Damascus, gazId + doc/<id>.json full record); terms good — CC-BY 4.0 stated in the site footer; structure excellent — robots.txt 'User-agent: * / Allow: /', JSON endpoints search.json, doc/<id>.json, suggestions, children, geoSearch; coverage fit good — populated-place, archaeological-site and building-institution types with coordinates and polygons across the Islamic-period Mediterranean and Near East; corroboration good — per-record identifiers (zenon-systemnr, zenon-thesaurus, GND-ID) and owl:sameAs links to Pleiades/geonames, the same identifier families the museum's place records already carry as external_ids; language coverage good — names in about 25 languages including Arabic; longevity excellent — institutional DAI service, versioned releases, per-record provenance batches. dec-20261109-228 · 2026-11-09
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 2 citable family, facts-only: cite {cite:idai-gazetteer <gazId>} for place-identity facts (canonical place URI, gazetteer types, coordinates, alternate names including Arabic, parent/ancestor hierarchy); gazetteer editorial text never republished; harvest via search.json and doc/<id>.json endpoints; entity enrichment limited to an idai external id plus source citation after per-record identity verification
dec-227 docked a source-onboarding question at the governance gate because the literal string 'robots.txt' appears in the access-posture evidence — one of the nd-governance keyword surfaces. The gate fired on a lexical mention, not on any actual governance change: no schema, enum, predicate, route, header, infra or CI surface is touched; the question asks only whether a public scholarly gazetteer joins the citation families, the same class as the already-decided web-* onboarding questions. The correct ruling applies the eight onboarding criteria in docs/agents/source-families.md: the German Archaeological Institute is the federal authority for classical archaeology and the gazetteer is a Pelagios linked-data standard; canonical place/<gazId> URIs and doc/<id>.json records are stable and machine-readable; the content licence is a real open licence (CC-BY 4.0, stated in the site footer); robots policy allows all agents and the JSON endpoints answer anonymously; coverage spans the Islamic-period Mediterranean and Near East with coordinates, polygons and ~25-language names including Arabic; owl:sameAs links to Pleiades/geonames corroborate identifier families the place records already carry. Tier 2 rather than Tier 1 because the gazetteer is a shared reference vocabulary for place identity — the same role Nomisma plays for numismatic concepts — while Tier 1 is reserved for primary institutional record databases. Facts-only regardless: gazetteer editorial text is never republished.
Should the museum mark a canon of cornerstone and major records with an optional `importance` field (1 or 2) on the entity, selected by a recorded rule (graph weave plus Wikidata sitelink reach) with a short editorial include list, so that finishing, featuring and search ranking can key off it? dec-20261110-229 · 2026-11-10
طے شدہ · rule · لین: canon · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: modern figures are admitted by the lane's reach threshold — a published present-day reach figure from a citable source; below threshold the name stays a queue lead, not a record
Answered by a standing repository rule (rule-reach-threshold): Method page: 'reach thresholds for modern figures'.
Should the museum mark a canon of cornerstone and major records with an optional `importance` field (1 or 2) on the entity, selected by a recorded rule (graph weave plus Wikidata sitelink reach) with a short editorial include list, so that finishing, featuring and search ranking can key off it? dec-20261110-230 · 2026-11-10
طے شدہ · override · لین: canon · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Add an optional `importance` field (1 cornerstone, 2 major) to the base schema and set it on 824 records by scripts/report/canon.py: eligibility needs evidence both from the museum's own graph (inbound edges, prose links, pathway appearances) and from Wikidata sitelink reach, ranked within per-class quotas, plus 59 named editorial includes at importance 2. The field ranks editorial priority only; it does not judge a subject's standing and changes nothing for records without it. Shia-tradition subjects that qualify on the same signals (for example Ja'far al-Sadiq, al-Kulayni, al-Kafi, the Imam Ali shrine, Najaf, Karbala) are kept, since the canon orders what to finish and the scope rule already governs how each page is framed; this is flagged for the founder to confirm or strike.
dec-229 answered with rule-reach-threshold, the Modern Islam Browser admission rule, because the question contains the word "reach". That rule decides whether a modern figure gets a record at all; it does not bear on marking priority among existing records. The founder approved the direction in chat; the design choices left to the agent (field on the entity rather than a side list, two levels, the selection rule, the include list) are recorded here. All are reversible and invisible to visitors until a later change uses the field.
The Place wing map was rebuilt as a plain script, so the Svelte component MapViewer.svelte is no longer drawn on any page. Taking it out of the build changes the hashed names of the script chunks it shares with the depth dial on every entity page, and so the HTML of about 160,000 pages. Should the unused component stay a build entry for now and be taken out in a separate change the founder can batch with other rewrites-all-pages changes? dec-20261111-231 · 2026-11-11
طے شدہ · rule · لین: place-map · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: media only from schema/enums/licenses.yaml; third-party content keeps its own licence; no figurative depictions of prophets or Companions; no AI-generated portraits
Answered by a standing repository rule (rule-media-license): AGENTS.md (licences rule).
The Place wing page carried 14 MB of map data inside its HTML (25.9 MB per language). Should the map script fetch that data on demand from static JSON files written beside the page under /<lang>/place-map/ at build time, instead of the page carrying it or a new /api/v1/ export serving it? dec-20261111-232 · 2026-11-11
طے شدہ · rule · لین: place-map · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: every feature is graph + API + search: records stay in the one graph, the class is exported in the static API, pages are search-indexed, and the feature is registered in docs/site/browsers.md, the plan index, the roadmap, docs/agents/lanes.md and docs/status.md — label anything submission-impacting
Answered by a standing repository rule (rule-feature-graph-api-search): AGENTS.md (every feature is graph, API and search) + docs/ai-change-gates.md §3.
On the Place wing map, 867 of 3,875 places could not be reached because they carry none of the twelve region ids, and 14 place records carry the coordinates [0, 0]. How should the map show them without changing any record: assign a display region by geometry and leave the [0, 0] places off the map but in the index, or leave them out, or edit the records? dec-20261111-233 · 2026-11-11
طے شدہ · rule · لین: place-map · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: media only from schema/enums/licenses.yaml; third-party content keeps its own licence; no figurative depictions of prophets or Companions; no AI-generated portraits
Answered by a standing repository rule (rule-media-license): AGENTS.md (licences rule).
The Place wing map was rebuilt as a plain script, so the Svelte component MapViewer.svelte is no longer drawn on any page. Taking it out of the build changes the hashed names of the script chunks it shares with the depth dial on every entity page, and so the HTML of about 160,000 pages. Should the unused component stay a build entry for now and be taken out in a separate change the founder can batch with other rewrites-all-pages changes? dec-20261111-234 · 2026-11-11
طے شدہ · override · لین: place-map · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Keep MapViewer.svelte as a build entry that no page draws, and take it out, with the three packages only it uses, in a separate small change labelled rewrites-all-pages for the founder to batch. The wing page holds one never-true expression that names the component, with a comment that says why.
dec-231 answered with the media licence rule, which does not bear on a build question. The choice is between an unused 126 KB script that no reader downloads and re-uploading about 160,000 pages for a change no reader can see; the house rule on deploy size settles it. Measured, not assumed: two scoped production builds at the same path give identical chunk names and byte-identical entity pages. Reversible by deleting the component and the expression.
The Place wing page carried 14 MB of map data inside its HTML (25.9 MB per language). Should the map script fetch that data on demand from static JSON files written beside the page under /<lang>/place-map/ at build time, instead of the page carrying it or a new /api/v1/ export serving it? dec-20261111-235 · 2026-11-11
طے شدہ · override · لین: place-map · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Static JSON files under /<lang>/place-map/, written at build time by site/src/pages/[lang]/place-map/[file].json.ts from the one graph: places.json for every place, and 2 x 128 small files found by hashing a place's slug or a record's id, fetched when a reader opens one. No /api/v1/ file, Worker, wrangler or workflow file changes. The places themselves stay in the graph, the class exports and search exactly as before, so the graph, API and search rule is already met. Under docs/ai-change-gates.md section 3, which names a script and a network call, the change is labelled submission-impacting and waits for the founder's sign-off; the reserved outreach files are edited only after it. No filed claim changes (same-origin static files, nothing stored, no third party).
dec-232 answered with the graph, API and search rule. That rule is satisfied and is not the question: the question is how a page's own script gets the data it draws. The standing example is timeline-data (page data moved out of a page into a static file), and the language path is already served from R2, which keeps the change inside the page's own files. Files are hashed by name so that a new record changes one file. Reversible: the endpoint and the script are the only parts that know the layout.
On the Place wing map, 867 of 3,875 places could not be reached because they carry none of the twelve region ids, and 14 place records carry the coordinates [0, 0]. How should the map show them without changing any record: assign a display region by geometry and leave the [0, 0] places off the map but in the index, or leave them out, or edit the records? dec-20261111-236 · 2026-11-11
طے شدہ · override · لین: place-map · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: Display only, no record edited. A place's region on the map is its first tag among the twelve regions, else the region whose outline holds it (tried up to two degrees around, because the 1:110m coastline cuts off islands and harbours), else Rest of the world. A place whose coordinates are [0, 0] is treated as having none: it is not plotted, and it stays in the index under the map. The 14 such records are listed in docs/site/place-map-review.md for their lanes to correct.
dec-233 answered with the media licence rule, which does not bear on how a marker is drawn. Leaving 867 places unreachable hides a fifth of the wing; editing 600 records is outside this task and is other lanes' work; a display rule in one library file fixes what a reader sees today and changes no data. Plotting [0, 0] would put Jabal Amel and a Cairo cemetery in the Gulf of Guinea, which is a false statement on a map. Reversible in site/src/lib/place-map.ts.
Hilyat al-Awliya vol. 10 notices 459, 460 and 648 (Ahmad ibn Khidruya al-Balkhi, Ibrahim al-Harawi known as Sitanba, Abu Amr al-Zajjaji) have no person record or alias in data/person. Create new draft person records from the single Hilyat notice, with no faith or tradition value because the notice states none, and read the garbled OCR nickname of notice 460 as Sitanba? dec-20261112-237 · 2026-11-12
طے شدہ · rule · لین: microsd-books · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only on two independent sources attesting it for the person himself; office, headship, descent, movement membership, infobox-only values, teacher lineages and aggregator rows never count; where sources say only 'Shia' and context supports a branch, the founder's mapping gives shia-twelver documented with the mapping stated — otherwise unknown
Answered by a standing repository rule (rule-faith-two-source): Method page doctrine + founder Shia-mapping rule 24 Oct 2026 + founder re-review 1 Nov 2026 ('strict two-source rule restored; branch-only attestation settles nothing').
Hilyat al-Awliya vol. 10: (a) notice 645 names Sa'id ibn Abd al-Aziz al-Halabi, a companion of al-Sari al-Saqati who settled in Damascus, while the graph already holds person/said-ibn-abd-al-aziz-al-tanukhi (the earlier Damascene jurist) — merge or keep apart; (b) notice 650 says Abu Bakr ibn Abi Sa'dan 'followed al-Shafi'i' — record that as a traditions/faith value on the new record or state it in cited prose only? dec-20261112-238 · 2026-11-12
طے شدہ · rule · لین: microsd-books · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only on two independent sources attesting it for the person himself; office, headship, descent, movement membership, infobox-only values, teacher lineages and aggregator rows never count; where sources say only 'Shia' and context supports a branch, the founder's mapping gives shia-twelver documented with the mapping stated — otherwise unknown
Answered by a standing repository rule (rule-faith-two-source): Method page doctrine + founder Shia-mapping rule 24 Oct 2026 + founder re-review 1 Nov 2026 ('strict two-source rule restored; branch-only attestation settles nothing').
Sensitivity label for a lecture-series record: work/silsilat-ebadu-al-rahman (Omar Abdelkafy, Assabile) episodes 9, 10 and 12 hold the lecturer's own opinions on disputed matters (abortion, contraception, women's fields of work and separate schooling, censorship, atheism called a sickness, an unsourced claim about imported food). Set sensitivity: contested on the record and mark each such point as the lecturer's view, per the audio-video harvest method rules 2 and 3 of 25 Oct 2026? dec-20261112-239 · 2026-11-12
طے شدہ · rule · لین: omar-abdelkafy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Hilyat al-Awliya 9/318 dates a conversation of Ali ibn Bakkar to a year whose units word is garbled in the OCR ('سنة ... ومائتين'), and one narrator's name on the same page is unreadable. Enter a floruit or death date and a reconstructed narrator name on the new record, or enter no date and describe the gap in the prose? dec-20261112-240 · 2026-11-12
طے شدہ · rule · لین: microsd-books · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Hilyat al-Awliya vol. 10 notices 462 (Abu Turab al-Nakhshabi, the transmitter of Hatim al-Asamm's teaching; the notice gives no personal name and no death year) and 649 (Muhammad ibn Ulayyan al-Nasawi, whose notice is short and all related by al-Sulami through one intermediary): create draft person records from these single notices, with roles [mystic] and no faith or tradition value because the notices state none? dec-20261112-241 · 2026-11-12
طے شدہ · rule · لین: microsd-books · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: a person's faith is recorded only on two independent sources attesting it for the person himself; office, headship, descent, movement membership, infobox-only values, teacher lineages and aggregator rows never count; where sources say only 'Shia' and context supports a branch, the founder's mapping gives shia-twelver documented with the mapping stated — otherwise unknown
Answered by a standing repository rule (rule-faith-two-source): Method page doctrine + founder Shia-mapping rule 24 Oct 2026 + founder re-review 1 Nov 2026 ('strict two-source rule restored; branch-only attestation settles nothing').
Sensitivity label for a lecture-series record: work/al-dar-al-akhera (Omar Abdelkafy, Assabile) episodes 1 and 3 hold the lecturer's own opinions on disputed matters (amulets and charms called forbidden, women visiting graves, fixed condolence gatherings and the fortieth-day memorial called innovations, and a curse invoked on al-Hajjaj ibn Yusuf). Set sensitivity: contested on the record and mark each such point as the lecturer's view, per the audio-video harvest method rules 2 and 3 of 25 Oct 2026? dec-20261112-260 · 2026-11-12
طے شدہ · rule · لین: omar-abdelkafy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
A place record's own sources give no coordinates for it (its Wikidata item and Wikipedia articles carry none, and GeoNames has no entry). May the record take the coordinates of another feature, and under what condition? dec-20261112-261 · 2026-11-12
طے شدہ · rule · لین: place-coordinates · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: never guess an external id: if the QID or identifier cannot be confirmed from a live source, omit the field and state the gap in provenance — the wave-1 fail-closed precedent
Answered by a standing repository rule (rule-external-id-failclosed): wave-2 commit: 'the gap is stated in the provenance rather than guessed'.
A place record's own sources give no coordinates for it (its Wikidata item and Wikipedia articles carry none, and GeoNames has no entry). May the record take the coordinates of another feature, and under what condition? dec-20261112-262 · 2026-11-12
طے شدہ · override · لین: place-coordinates · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: A place record may take another feature's coordinates only when a cited source places that feature inside the record's place, so that the point lies on the place itself. The point is recorded with coordinate_precision approx, a coordinate_note that names the feature and says the point is not the place's measured centre, and a ledger page for each Wikipedia or Wikidata revision read. Applied once: place/al-wardya-cemetery takes the point of the mosque of Shaykh Umar al-Suhrawardi, which stands inside the cemetery. Where the sources name only a containing city or quarter, or a nearby feature, no coordinates are recorded by the agent: place/fadl-mosque, place/grave-of-al-ghazali, place/al-nawbakhti-shrine and place/sirwan keep their records unchanged and are listed for the founder in docs/delegation/place-coordinates-held-proposal.md, with city-level placement as an option for him to approve.
dec-261 answered with rule-external-id-failclosed, which is about omitting an identifier that cannot be confirmed. Its spirit (state the gap, do not guess) agrees with holding the four records that have no citable position, but it does not answer whether a point that a cited source places inside the record's place may stand for it. A point inside the cemetery is a true statement about where the cemetery is, at the precision the note declares; a city centre is not a statement about where a mosque or a grave is, and four Baghdad sites stacked on one pin would read as a position no source gives. The golden set does hold records placed at a regional centre, so that option is not ruled out: it is left for the founder. The change is one edit on one record and is undone by restoring the previous coordinates.
Hilyat al-Awliya notice 466 is Ali al-Jurjani, an early worshipper who met al-Sari al-Saqati on the road to Abbadan. data/person already holds al-sharif-al-jurjani-ali-ibn-muhammad (the author of the Definitions) and al-jurjani-ali-ibn-abd-al-aziz. Keep the notice's worshipper as a new record apart from both, or merge into one of them? dec-20261112-264 · 2026-11-12
طے شدہ · safe-default · لین: microsd-books · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Sensitivity label for a lecture-series record: work/ahl-al-hikma (Omar Abdelkafy, Assabile) episode 4 holds the lecturer's own opinions on disputed matters (fixed condolence gatherings, birthday and wedding anniversaries, engagement rings and marriage customs criticised as customs or imitation, and claims about the effects of missing the dawn prayer). Set sensitivity: contested on the record and mark each such point as the lecturer's view, per the audio-video harvest method rules 2 and 3 of 25 Oct 2026? dec-20261112-265 · 2026-11-12
طے شدہ · rule · لین: omar-abdelkafy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Should Papyri.info (the documentary papyrology portal of the Duke Collaboratory for Classics Computing, joining the Duke Databank of Documentary Papyri texts with the Heidelberg HGV metadata, stable DDbDP and Trismegistos identifiers, a public data repository under Creative Commons Attribution 3.0) be onboarded as a citable source family for dated documents of early Islamic Egypt, and at what tier? Assessment 11 Nov 2026 against the eight criteria of docs/agents/source-families.md: authority excellent (Duke DC3 with the Heidelberger Gesamtverzeichnis and editorial boards per component; the field's standard corpus); citability excellent (DDbDP hybrid ids such as sb;6;9576 resolving at papyri.info/ddbdp/<id>, plus Trismegistos numbers); terms good (the data repository README states Creative Commons Attribution 3.0 with attribution to the respective projects); structure good (EpiDoc XML per document in the public data repository, synchronised hourly; the live site answers scripts with a bot challenge, so the repository is the only harvest channel); coverage fit moderate (Greek and Coptic documents of first-century-AH Egypt: receipts, requisitions and governors' correspondence naming officials, places and dates, feeding place and person records for Egypt and the Fustat everyday-life gap; most of the corpus is pre-Islamic and out of scope); corroboration good (contemporary dated documents, the earliest written evidence for the new administration, with HGV provenance tied to Pleiades and Trismegistos place ids); language coverage weak for Arabic (texts Greek and Coptic, metadata German; the Arabic documents themselves stay with the already registered Arabic Papyrology Database); longevity excellent (institutional hosting, version-controlled public data, identifiers independent of URLs). dec-20261112-266 · 2026-11-12
طے شدہ · override · لین: web-sources · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Tier 2 citable family, facts-only: register source/papyri-info. Cite {cite:papyri-info <ddb-id>} for documentary facts (document type, date, provenance with HGV certainty, parties named) in the museum own words. Harvest only from the public data repository (DDbDP and HGV_meta_EpiDoc EpiDoc XML), never from the challenged live site. Edition text and translations never republished.
The two engine decisions came from lexical matches on words in the assessment ("verified", "source family"), not from any governance, review-status or independence question. The onboarding question itself is the same class as the registered Tier 2 document families. The eight criteria in source-families.md are met on the repository evidence (authority, citability, CC BY 3.0 terms, structure via the public repository, coverage of early Islamic Egypt documents, corroboration, longevity). Arabic-first value is weak (texts are Greek and Coptic) and is recorded as such. Facts-only regardless.
Sensitivity label for a lecture-series record: work/ajaeb-al-kulub (Omar Abdelkafy, Assabile, 29 short Ramadan episodes on diseases of the heart) holds the lecturer's own opinions and unchained reports on disputed or practical matters (diet and weight in Ramadan, remedies and protective recitations, and, in later episodes, social and political remarks). Set sensitivity: contested on the record and mark each such point as the lecturer's view, per the audio-video harvest method rules 2 and 3 of 25 Oct 2026? dec-20261112-267 · 2026-11-12
طے شدہ · rule · لین: omar-abdelkafy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Sensitivity label for a lecture-series record: work/mawkib-al-nour (Omar Abdelkafy, Assabile, daily Ramadan programme on one or two verses a day) holds the lecturer's own opinions and unchained reports on disputed or practical matters (diet and weight in Ramadan, dress of women in a household, lay remarks on psychology, blessing in provision and the dawn prayer). Set sensitivity: contested on the record and mark each such point as the lecturer's view, per the audio-video harvest method rules 2 and 3 of 25 Oct 2026? dec-20261112-268 · 2026-11-12
طے شدہ · rule · لین: omar-abdelkafy · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: harvest inclusively: any intelligible topic is included, contested windows are included as the speaker said them, opinions/rants are categorised by closest fit, garbled windows are discarded and reported — nothing silently dropped
Answered by a standing repository rule (rule-harvest-include): founder harvest rulings 25 Oct 2026 (docs(harvest) commits).
Build and deploy speed changes (graph lookups, non-blocking R2 staging upload, packed build artifact, no shared status file): the founder said in chat on 12 November 2026 to leave production as it is, try the changes on UAT first and make them on release/uat rather than main, although AGENTS.md (founder rule of the same day) retires release/uat. After the UAT build and deploy of 13 November he said to open the pull request to main and to fix the /ar redirect and the action versions with it. Which branch applies to the trial, and may the changes then go to main? dec-20261113-269 · 2026-11-13
بانی کے پاس · policy-gate · لین: build-deploy-performance · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
Hosting fallback: the documents name a GitHub Pages fallback mirror for a Cloudflare outage, but no workflow publishes it (last GitHub Pages deployment 18 October 2026) and GitHub Pages cannot hold the site (1 GB limit; the site is about 19 GB and 456,000 files). What is the fallback for a Cloudflare outage? dec-20261113-270 · 2026-11-13
بانی کے پاس · policy-gate · لین: build-deploy-performance · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: deploys are founder-only (cost + control)
CI flow for pull requests (docs/engineering/46-ci-flow-proposal.md): the founder approved the proposal as written in chat on 13 November 2026. Its implementation left choices to the agent: how large the sample of record pages is, whether the release notes bump the site version, and whether the new full link check of main may fail a build run. What applies? dec-20261113-271 · 2026-11-13
بانی کے پاس · founder-reserved · لین: build-deploy-performance · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
Prophets' Tree on the Messengers Browser: which lines may be drawn as lineage when the graph records only a few father-and-son edges between the prophets? dec-20261113-272 · 2026-11-13
طے شدہ · override · لین: messengers · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Draw only what the graph holds as lineage: a solid branch for a child_of edge, the Ismail-to-Muhammad vine from ANSAB_PROPHET_CHAIN with its disputed part dashed, a dashed twig for a prophet the detector finds named in the cited ayah of al-An'am 84 to 86, and every other place marked as order of mention. checkMessengersTree fails the build when a drawn line has no edge. The classical chart's fuller lineages enter the tree when they enter the graph with their sources.
No rule or precedent addressed how a drawn lineage may relate to the graph, so the engine held at safe-default. The Founder's Charter makes the graph the foundation and the site its application, and AGENTS.md forbids inventing a fact or citation; a lineage typed into a layout would be both. Reversible: a branch is promoted by adding the sourced edge.
Signature views of the browsers (docs/site/signature-views.md): how does a view treat a record that has no date, no coordinates or no edge, when the drawing places records by date, place or relation? dec-20261113-273 · 2026-11-13
طے شدہ · override · لین: signature-views · اعتماد: 1 · الٹنے کے قابل
فیصلہ: A record with no date, coordinates or edge is left off the drawing, counted in the view's caption and kept reachable from the lists below it. No view estimates a date or position.
The engine held at safe-default: no rule states what a view does with a record it cannot place. Placing it at a guessed date or position would invent a fact (AGENTS.md, Sources). Dropping it from the page would hide a record. Counting it keeps the drawing true and the record findable. Stated as rule 2 of docs/site/signature-views.md.
Mukatabat Browser signature view: the design called for a map of letters with an arc from the sender's place to each recipient's place. The curated letter records carry only an occurred_at place and had_participant edges, with no pair of sender place and recipient place. Draw the map or not? dec-20261113-274 · 2026-11-13
طے شدہ · override · لین: signature-views · اعتماد: 1 · الٹنے کے قابل
فیصلہ: No map of letters is drawn. The Mukatabat view is the flow from sender class to kind to corpus only, and the register records the omission. The map becomes possible when the letter records carry both a sender place and a recipient place.
The engine answered with rule-media-license, matched on the words map and draw; the question is about whether the records support arcs between places, not about media or licences. Drawing arcs to recipients' seats from general knowledge would state places the records do not hold. Reversible: the map is added when the data exists.
Ghayb Browser, the ascent: journey/isra-and-miraj-route holds two waypoints in order (Mecca, then Jerusalem). The page's old code followed them with three heavens concepts (the seven heavens, Sidrat al-Muntaha, al-Bayt al-Ma'mur) in a fixed order of its own. May the view number those three as steps of the ascent? dec-20261113-275 · 2026-11-13
طے شدہ · override · لین: signature-views · اعتماد: 1 · الٹنے کے قابل
فیصلہ: Only the journey record's own waypoints are numbered on the ascent. The three heavens concepts are shown after them, unnumbered, as concepts of the browser's heavens section.
The engine held at safe-default. The order of the three concepts was a constant in the page's code and is stated by no record; numbering them as steps of the ascent would assert a sequence without a source. The journey record's waypoints carry their own order, so they keep their numbers.
Ibadat Browser rosette: the practice records carried no worship family, and the view grouped 41 of them from a table typed into the component. Where should the grouping live, and what happens to a practice whose record names no family (for example practice/nadhr, which had been placed under supplication only through the Du'a Browser's section tag)? dec-20261113-276 · 2026-11-13
طے شدہ · override · لین: signature-views · اعتماد: 1 · الٹنے کے قابل
فیصلہ: The worship family of a practice is an ibadat-<topic> tag on the practice record, read by ibadatTagTopics; 41 records carry it and the table in the component is removed. A practice whose record names no family is listed under the rosette, not placed; eleven await classification from sources.
The founder instructed in chat on 13 November 2026 to add a tag to each practice record and remove the table. The engine held at safe-default, having no rule for where a view's grouping lives. practice/nadhr had been placed under supplication only through the Du'a Browser's section tag, which is not a statement about its worship family, so it joins the unplaced row until its record names one.
Which next steps for lanes A, B and C did the founder approve in chat on 13 and 14 November 2026, and which questions stay held? dec-20261114-277 · 2026-11-14
بانی کے پاس · founder-reserved · لین: orchestrator · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
How should lane workers handle a blocker or a pending decision: stop and wait, or keep moving? dec-20261114-278 · 2026-11-14
طے شدہ · safe-default · لین: orchestrator · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Register the Surah web app (web.surahapp.com, publisher Tafsir Center for Quranic Studies, public API at dev.surahapp.com) as a facts-only source family for translation, commentary catalogue and metadata facts; the scripture wording itself is never taken from it. No terms page found; robots.txt sets no content signals; API open and documented. Should it be registered facts-only? dec-20261114-279 · 2026-11-14
بانی کے پاس · policy-gate · لین: web-sources · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: Qur'an text is never agent-decided
Should person/rawh-ibn-hatim be merged into person/ruh-ibn-hatim-al-muhallabi (same man, died 174 AH, al-Zirikli al-A'lam vol. 3 p. 34), keeping the richer record person/ruh-ibn-hatim-al-muhallabi, with the other slug kept as an alias and every reference repointed? dec-20261114-280 · 2026-11-14
طے شدہ · safe-default · لین: web-sources · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Gemini returned verdicts on the pending-corrections ledger (14 Nov 2026). Should its sections 1 to 5 be applied only where each claim is independently confirmed against a primary source read by the worker, and left unchanged and logged as unresolved otherwise, with section 6 kept as advice for the founder? dec-20261114-281 · 2026-11-14
طے شدہ · rule · لین: corrections · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: corrections are made openly: fix the record, cite what the correction rests on, append a CANARY entry — errors are corrected when found, never hidden
Answered by a standing repository rule (rule-corrections-open): AGENTS.md + Method page: 'corrections in the open'.
Should one central file in git, docs/library/harvest-register.yaml, record every book, audio series, website and dataset harvested, OCRed, mined or ingested, so that every agent looks an item up before any download, OCR, harvest, mining or ingestion, and a finished book's PDF and working files may later be deleted from the microSD, with its OCR text kept as one zip on the microSD? dec-20261114-282 · 2026-11-14
بانی کے پاس · policy-gate · لین: library · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: destructive operations are never autonomous — negation tricks must not disarm this gate
The founder instructed in chat on 14 November 2026: "we also need a register of all the pending records, all the records that are missing wikidata id or anything else you or other agents still wants to enrich, and all the proses that are partial or empty. and their status, so that at any time any agent can directly and easily find those and work on them." Should the museum keep an enrichment register under docs/library/enrichment/ (a script that derives every gap live from data/ and content/, sharded queue files, an append-only claims.yaml, and a check script), with the AGENTS.md rule that agents take work from it when their queue is empty, claim it in their first commit and never start an item another agent has claimed? dec-20261114-283 · 2026-11-14
بانی کے پاس · policy-gate · لین: library · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: governance surfaces (schema, classes, enums, predicates, routes, headers, infra, CI) are founder-only
The founder approved four changes in chat on 14 November 2026: "yes make the changes 1 to 4 that you proposed". (1) Release notes: lane workers write one release-note entry per packet (the manifest's canary field) and the lander puts them into CANARY.md in the same pull request, so "every change has its entry in the same commit" still holds. (2) Batch size: the limit of 25 entities applies per packet, and one train pull request may carry many packets. (3) Validation: a fast scoped check per packet (npm run check:packet), and the full npm run validate once per train before the merge. (4) Who merges: only the lander merges lane content; lane workers no longer open or merge their own pull requests. Reason: twenty agents merging small pull requests conflicted on CANARY.md and the progress files, and output stayed flat. Should AGENTS.md and the agent guides be amended to record the packet pipeline as the standing procedure for lane content, leaving work outside the lanes (site code, docs, fixes) on the existing workflow and weakening no content rule? dec-20261114-284 · 2026-11-14
بانی کے پاس · founder-reserved · لین: pipeline · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
The founder ruled in chat on 15 November 2026: "make the changes and adjust the claims and be honest on everything". The public-claims audit of the same day (docs/audit/public-claims-audit-2026-11-15.md, merged in PR #7045) measured the museum's public statements against its own records and found the draft Museums + Heritage Awards entry overstated AI's role as mechanical-only, implied a book-only corpus and a human-reviewed published tier, and used stale cost and count figures. The founder's ruling: every public claim about the museum — in award entries, submissions, outreach and the site's own pages — must match what the museum's own public record shows: AI tools research, draft, translate and check entries; most records are drafts; sources include reference sources and records whose sources are not yet checked, stated as such; running costs are the site's own published figures; and no text may imply a human-reviewed corpus. Should this honesty rule be recorded as a standing founder ruling, binding on every public statement about the museum, with the audit ledger as its evidence and the awards entry of 15 November 2026 as its first application? This question is reserved to the founder: the founder decides. dec-20261115-285 · 2026-11-15
بانی کے پاس · founder-reserved · لین: orchestrator · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
The founder ruled in chat on 15 November 2026: "for the ai-change-gates, do not require a human to read, so anything can be published state if the ai lists its sources clearly and if the sources are difficult then mention that on the record itself and instead of draft say reader can verify and that the website has done its job in attempting to collect and present that material to the user, in the future we might be able to run other cycles and verify from other resources for that record." This supersedes the founder ruling of 8 November 2026 (decision records dec-20261108-152/161, restated in AGENTS.md) that AI output is draft-tier until a human reads it. The new rule: no human read is required for a record to stand as published; a record may be published when the AI lists its sources clearly; where sources are difficult, that difficulty is stated on the record itself; the status language is reframed from draft-as-awaiting-a-human to reader-can-verify — the website has done its job in attempting to collect and present the material, and future cycles may re-verify a record from other resources. Should the publication rule be amended accordingly — docs/ai-change-gates.md section 6 and docs/agents/publication-standard.md aligned to the ruling's full terms, and the 8 November ruling in AGENTS.md marked superseded? This question is reserved to the founder: the founder decides. dec-20261115-286 · 2026-11-15
بانی کے پاس · founder-reserved · لین: orchestrator · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: the question reserves this decision to the founder
Register Oxford Islamic Studies Online (as archived by the Wayback Machine) and, as a general family, defunct sites preserved by the Internet Archive, as trusted source families for facts-only enrichment? dec-20261116-287 · 2026-11-16
طے شدہ · rule · لین: oxford-islamic-studies-wayback · اعتماد: 0.9 · الٹنے کے قابل
فیصلہ: independence is counted by source FAMILIES, not documents: ten mirrors or republications of one family are ONE attestation, not ten; two attestations require two independent families; a copied single-source 'second source' never counts
Answered by a standing repository rule (rule-source-families): Muse doctrine §5 (independence is source families) + faith two-family rule.
Jami al-Tawarikh source records (source/jami-al-tawarikh-khashshab-1960 and source/tarikh-wassaf) need a cited source for their non-open layer under the validator. May a source record cite the scanned edition itself (the archive.org scan and the Khashshab scan, layer copyrighted-reference), or must it cite an external bibliographic record? dec-20261122-297 · 2026-11-22
طے شدہ · safe-default · لین: jami-tawarikh · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Jami al-Tawarikh (Rashid al-Din, Ilkhanid history, volume 2 part 1, scan pages 1-60): the record for the 700 AH execution of the Alid naqib Taj al-Din Abu al-Fadl Muhammad, and the two open questions on Rashid al-Din's birth year and claimed Jewish origin, cannot carry the museum's perspectives block because the validator requires a Sunni-tagged first view and the book and its sources give none. May the record state only the book's own account, attributed to the book, and mark the missing Sunni view as a source gap? dec-20261122-298 · 2026-11-22
طے شدہ · rule · لین: jami-tawarikh · اعتماد: 0.85 · الٹنے کے قابل
فیصلہ: contested topics use the perspectives block: Sunni position(s) first, then the academic view, then any other view — each cited and labelled, no verdict in the museum's voice
Answered by a standing repository rule (rule-perspectives): AGENTS.md (voice and scope).
Cost: should the 10-minute pipeline check and land cycle be lengthened to 30 minutes? Most land cycles now find nothing to land, and each run costs a full coordinating turn. dec-20261122-310 · 2026-11-22
بانی کے پاس · policy-gate · لین: assabile-pipeline · اعتماد: n/a · الٹنے کے قابل نہیں
فیصلہ: review-status escalation is founder/pipeline only
The Jami al-Tawarikh scan we hold is volume 2 (parts 1 and 2). Volume 1 is not in the archive.org item. Should the search for volume 1 continue, or should the book be mined as volume 2 only? dec-20261122-311 · 2026-11-22
طے شدہ · safe-default · لین: jami-tawarikh · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Cost: should the coordinating session (the one that runs the 10-minute check and land cycles) switch from Sonnet 5.5 to Haiku 5.5, given that all mining and worker agents already run on Haiku 5.5? dec-20261122-312 · 2026-11-22
طے شدہ · safe-default · لین: assabile-pipeline · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Worker W3b has been on one episode (el-Sabaa el-Mubikat ep 03) for well over two hours with a fresh heartbeat, but has not finished it. Restart it on another file, or let it continue? dec-20261122-313 · 2026-11-22
طے شدہ · safe-default · لین: assabile-pipeline · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔
Worker W6 (binbaz.org.sa fatwa crawl) has restarted about nine times in ninety minutes, each time hitting the same unrecognised page (id 1832) and backing off. Should the crawler be restarted again, or stopped until a parser or site check is done? dec-20261122-314 · 2026-11-22
طے شدہ · safe-default · لین: assabile-pipeline · اعتماد: 0.4 · الٹنے کے قابل
فیصلہ: hold as draft-tier / pointer-only; record question in lane progress file for the next precedent-extraction pass
No rule or precedent matched. Defaulting to the most conservative reversible action per the decision-ledger design.
کوئی قاعدہ یا سابقہ میل نہیں کھاتا — جب تک سابقہ موجود نہ ہو، محتاط طے شدہ صورتِ حال برقرار ہے۔