faq-optimization

name	faq-optimization
description	Use when: kullanıcı "FAQ optimize et", "FAQ ekle", "FAQ enhance", "FAQ schema yenile", "AIO citation pattern uygula" der ya da `/pseo-faq-optimization` çağırır. master.xlsx[content_decay] action trigger ya da master.xlsx[content_improve] optimization trigger; existing blog FAQ section enhance (mode=enhance) veya yeni blog FAQ add (mode=add). R-09 talep-güdümlü FAQ (kanıt varsa 3-6, hard cap 10) + R-43 statik (accordion YASAK) + R-29 pasaj alıntılanabilirlik gate + R-79 FAQPage @graph schema + R-109/R-110/R-111 AIO citation pattern enforce. Also use when: existing blog FAQ section'da AIO citation pattern eksik detect edildiğinde (per 500 word density 1-2 dışı); pasaj alıntılanabilirlik gate (R-29) fail edildiğinde (cevap-önce TL;DR yok); FAQPage schema markup eksik veya outdated (R-79 @graph Question entity boş); project-config[profile] enum 5-değer (e-commerce | ymyl | local-service | b2b-saas | portfolio) içinden biri sabit ve FAQ taxonomy profile-aware enforce edilmeli (Principle 2). Do not use when: yeni blog için (new-blog skill kullan, FAQ inline rendered); content prune/redirect/delete için (content-remediation Wave 2 domain); image generation için (generate-images kullan); decay action="revise" full section için (revise-content kullan, R-87 section-targeted); existing HTML diskte yok mode=enhance (DURUR #1, FAQ section parse edilemez); R-09 cap aşımı 11+ FAQ attempt (DURUR #2, AI suistimal Principle 3 enforce).
version	1.0
status	wip
category	production
inputs	{"project_slug":{"type":"string","required":true,"description":"Workspace proje slug (resolves projects/{slug}/master.xlsx + outputs/blog/{slug}/)."},"url":{"type":"string","required":true,"description":"Existing blog URL (mode=enhance) veya target URL (mode=add)."},"mode":{"type":"string","required":false,"default":"enhance","description":"FAQ mode — 'enhance' (existing FAQ section update) veya 'add' (yeni FAQ section ekle). Brief enum [enhance, add]."}}
outputs	["outputs/blog/{slug}/article.html","_state/events.jsonl"]
consumes	["init-project:projects/{slug}/master.xlsx#new_content_plan","init-project:projects/{slug}/master.xlsx#content_decay","init-project:projects/{slug}/master.xlsx#content_improve","init-project:projects/{slug}/project.config.json","rules:rules/content-quality.md","rules:rules/content-html-discipline.md","rules:rules/content-seo-discipline.md","rules:rules/content-eeat-discipline.md","rules:rules/content-llm-discipline.md","rules:rules/content-update-discipline.md","templates:templates/content/faq-block.template.html"]
produces	["indexing-ping"]
triggers	{"manual":["/pseo-faq-optimization"],"natural_language":"\"FAQ optimize et ve AIO citation density per 500 word ayarla\",\n\"FAQ ekle ve R-09 talep-güdümlü 3-6 band hard cap 10 zorla\",\n\"FAQ enhance et ve R-29 pasaj cevap-önce TL;DR garantile\",\n\"FAQ schema yenile ve R-79 FAQPage @graph Question entity refresh\",\n\"AIO citation pattern uygula ve R-109/R-110/R-111 enforce et\"\n","hooks":[],"scheduled":[]}
mcp_tools	{"required":["mcp__gsc__search_analytics"],"optional":["mcp__dataforseo__serp_organic_live_advanced"]}
budget	{"uses_paid_mcp":true,"estimated_credits":3}
autonomy	{"confidence":"HIGH","requires_approval":true,"safe_auto_execute":false}

faq-optimization — production skill (Phase 11 Wave 2 W-F3)

FAQ optimize / enhance / add skill. master.xlsx[content_decay] action veya master.xlsx[content_improve] optimization trigger; existing blog FAQ section enhance (mode=enhance) veya yeni blog FAQ add (mode=add). R-09 FAQ count cap + R-43 statik HTML render (accordion YASAK) + R-29 pasaj alıntılanabilirlik gate + R-79 FAQPage @graph schema + R-109/R-110/R-111 AIO citation pattern enforce. READ-ONLY contract: master.xlsx 3 sheet (new_content_plan + content_decay + content_improve) sadece consume; sadece outputs/ + _state yazar.

Foundational Principles (Üst-Prensip — Alt-Rule Override Edemez)

Principle 1 — Truth-Verifiable Content (R-27, 3-katman defense)

Tüm FAQ content/source/link/data %100 doğru ve kanıtlanabilir; uydurma yasak (FAQ Q&A, source citation, expert quote, stat). FAQ context'inde:

• Layer 1 (pre-generate): skill prompt başlangıcında "uydurma FAQ yasak" sentinel; her FAQ Q&A için kaynak/source field zorunlu (no-source → reject).
• Layer 2 (post-generate): fact-check pass — her FAQ Q&A için kaynak verify (R-44 source verification 3-katman, R-105 expert quote bank-driven).
• Layer 3 (citation): citation enforce — eklenen her stat/quote source URL ile bağlanır, FAQPage acceptedAnswer.text içinde citation reference intact.
• Failure mode: P1 fact-check fail → DURUR #5 RED (FAQ section discard, yayın iptal, dateModified bump YOK).

Principle 2 — Profile-Aware Enforcement (project-config[profile])

Skill behavior project.config.json[profile] enum'una göre değişir. Enum 5-value: e-commerce | ymyl | local-service | b2b-saas | portfolio. FAQ context enforcement:

• profile == "ymyl" → R-51 disclaimer template ZORUNLU (medical/ legal/financial), counter-argument FAQ aday (R-50 cross-link); FAQ section başında veya per-FAQ disclaimer present check.
• profile == "e-commerce" → product FAQ taxonomy (return policy, shipping, sizing), conversational tone (brand_identity.tone consume).
• profile == "b2b-saas" → technical FAQ taxonomy (integration, pricing tiers, security, API docs); formal tone preserve.
• profile == "local-service" → location-specific FAQ (service area, hours, contact, NAP signal preserve).
• profile == "portfolio" → minimal FAQ (3-5 max, opsiyonel); R-09 cap 10 hard yerine 5 soft cap profile-driven downgrade.

Principle 3 — AI Suistimal Önlemi (Anti-Cheap-Content)

AI'ın doğal cheap content padding davranışını preempt et. FAQ context'inde (R-09 + R-29 + R-118):

• R-09 FAQ count cap (talep-güdümlü): kanıt varsa (PAA varlığı, gerçek kullanıcı soruları) 3-6 FAQ; kanıt yoksa daha az veya hiç; hard cap 10; 11+ FAQ attempt → DURUR #2 REJECT.
• Heading keyword density per FAQ %40-60: primary_keyword vs FAQ Q heading; aşırı stuff (>%60) AMBER, eksik (<%40) AMBER.
• R-29 pasaj alıntılanabilirlik gate: H3 cevap-önce + 50-150 word TL;DR per FAQ; first sentence izole parse self-contained.
• AIO citation density: per 500 word 1-2 citation (R-109/R-110/R-111); over-citation (>2 per 500w) DURUR #6 AMBER auto-correct.
• AI signature humanize: tone_phrases_blocklist consume (brand_identity), AI cliché phrases regex strip.
• Failure mode: AMBER warning (auto-correct attempt) → RED fail (manuel revise gerekir).

R-121 Bank Selection Logic (Conditional Applicability)

Conditional applicability is the headline. FAQ items are typically definitional (Q: "What is X?" / A: "X is a Y that does Z."). They often need no bank entry at all. R-121 fires for FAQ optimization only when a FAQ answer asserts an experience claim ("we measured ...", "we tested ...", "in our deployment ...") or a research claim ("our survey of 200 customers found ..."). For purely definitional FAQs, R-105 / R-114 / R-119 are not in scope, and R-121 has nothing to filter.

Cross-skill cap interplay (CRITICAL). R-121 density is per content, not per skill invocation. When mode == "enhance" and the existing article already cites bank entries in the body (originally placed by new-blog), those entries already consume the per-profile cap. faq-optimization MUST NOT add bank entries that would push the combined count over the cap.

Detection (mode == "enhance"): during Step 1 existing-blog parse, the skill also extracts current bank citations from the body (same two paths as revise-content — data-bank-entry-id attribute first, text-matching fallback for older articles). The result is bank_entries_pre_body[]. The FAQ pass treats this as already- committed cap consumption.

When R-121 DOES fire (FAQ answer asserts an experience/research claim), the same 3-step filter applies, with a FAQ-specific tightening on filter 2:

1. Topic match. entry.applicable_topics ∩ (new_content_plan. primary_keyword ∪ faq_question_text_keywords ∪ topical_map[matching _row].cluster) ≠ ∅ → entry remains a candidate; empty intersection → entry skipped.
2. Profile density cap, FAQ-tightened. The article's per-profile cap is the ceiling for the BODY + FAQ combined. FAQ-only additional cap, on top of any body usage:
   • YMYL: at most 1 experience entry inside the FAQ section (and the combined body+FAQ count still ≤ 2 experience + 1 research).
   • b2b-saas: 0 experience + 0 research inside FAQ by default (technical FAQ taxonomy rarely cites first-hand or original research — escalate to manual review if a FAQ item genuinely needs one).
   • e-commerce / local-service / portfolio: 0 inside FAQ by default (the body's 1 experience entry is usually sufficient).
3. Rotation (30-day). For each new candidate, count usage in master.xlsx[completed_work] rows where bank_entry_id == entry.id AND timestamp >= now - 30d. Skip if count >= entry.max_usage_per_month. Pre-existing body entries (extracted in mode == "enhance") are exempt from rotation — removing them only to satisfy the cap would be churn and would trigger a needless R-88 freshness-theater audit.

If all candidates fail the filter, the skill emits AMBER, retries with phrasings[] rotation; 2x AMBER same pass → RED upgrade (the FAQ section ships without that particular claim, or the operator manually reviews).

R-110 Anti-Pattern alignment. R-110 over-citation gate (>2 per 500w) and R-121 density cap reinforce each other: a FAQ answer that needs more than one bank entry to be credible is usually either too broad (split into separate FAQ items) or actually a body paragraph in disguise (move to article body via revise-content). The two rules together push FAQ items toward atomic, definitional Q&A structure.

Post-publish state mutation (Step 7 events.jsonl append): for each entry newly introduced in this FAQ pass, its last_used_in_content_id is set; the master.xlsx[completed_work] row carrying this content also records the bank_entry_id(s) added. Pre-existing body entries are not touched (already counted when new-blog or revise-content first shipped them).

Schema enablement (v1.4, commit 8e07e1c): R-121 reads applicable_topics, phrasings, last_used_in_content_id, max_usage_per_month on each bank entry. Stage C of brand-onboarding (commit cb8df43) populates these via R-44 evidence-gated atomic write.

Runtime integration deferred to Phase 11 Wave 2/3. This SKILL.md section is the spec lock; scripts/production/faq_optimization.py does not yet exist. When the runtime lands, the conditional applicability check (does this FAQ item assert experience/research?) runs first; only if YES does the 3-step filter execute.

Schema Authority Compliance

• event_kind: events.jsonl event_kind=work (events.schema.json enum 4-value: provenance | work | audit | workflow; ADR-020 production output → work).
• events.schema: events.jsonl 5 required field zorunlu (schema_version const "1.0", event_kind, event_id UUIDv4, timestamp UTC ISO 8601, project_id pattern ^[a-z][a-z0-9-]*$); event_type 12-enum içinden content_revise (FAQ enhance/add içerik revize semantiği — events.schema kanonik enum compliance).
• F-15: master.xlsx[content_improve] 8 col + allowed_writers null → READ-ONLY consume; W-F3 + W-F4 candidate input source (consistency report FAIL trigger için optimization sheet read).
• F-2 (Wave 1 reuse): master.xlsx[content_decay].action schema'da type/enum null → R-86/R-87/R-90/R-91 rule-derived consume; W-F3 sadece "revise" ve "manual FAQ enhance" branch consume eder, prune/ redirect/delete bu skill domain DEĞİL.
• F-6: master.xlsx 3 sheet (new_content_plan + content_decay + content_improve) allowed_writers null → READ-ONLY contract; transaction.append/update/delete YASAK.
• F-3 (Wave 1 reuse): project-config schema 1.2 + profile enum 5-value Wave 1 W-F1'de cascade fix uygulandı; bu skill profile field'ı var olduğu varsayımıyla consume eder (Principle 2 switch).

Routing (7-Step Workflow)

Step 1: master.xlsx Read + Mode Branch

input.mode switch:

• mode == "enhance" (default) → existing blog parse (outputs/blog/{slug}/article.html filesystem read).
  • Existing FAQ section detect (<section class="pse-faq"> veya <h2>FAQ</h2> selector; BeautifulSoup parse).
  • Current FAQ count + Q&A extract (h3 + p pair).
  • DURUR #1 trigger: existing blog YOK → impossible (R-89 canonical preserve enforce edilemez; new-blog skill'e yönlendir).
• mode == "add" → target URL find via master.xlsx[new_content_plan] row read where url_slug == {input.url} (14 col schema baseline: id, title, url_slug, primary_keyword, monthly_volume, assigned_cluster, target_word_count, priority, created_date, tivl_tag, lifecycle_status, image_prompt, alt_text, content_type).
  • new_content_plan kontrol: lifecycle_status == "GREEN" gerekli (yeşil ışık olmadan FAQ add yasak).

master.xlsx[content_improve] 8 col cross-check (existing_url, gsc_clicks, gsc_impressions, current_position, issue, optimization, target_position, priority): FAQ optimization trigger varsa optimization field'ı içinde "FAQ" keyword arar (READ-ONLY filter).

Step 2: project-config Read (Profile-Aware Switch — Principle 2)

projects/{slug}/project.config.json parse:

• profile field read (Principle 2 enum 5-value; F-3 schema 1.2 baseline).
• content_settings 14 field consume (toc_strategy, related_posts_strategy, author_strategy, css_strategy, indexnow_enabled, ai_training_optin, video_integration, internal_data_sharing, external_uniqueness_check, original_research_database, experience_database, video_database, disclaimer_templates, image_model).
• brand_identity 18 field consume (logo_url, primary_color, secondary_color, accent_color, font_family_heading, font_family_body, header_template_id, footer_template_id, source_url_for_sampling, tone, hitap, anglicism_tolerance, tone_phrases_blocklist, font_heading, font_body, default_hero_url, same_as_urls, image_style).

Profile-aware FAQ taxonomy switch:

• profile == "ymyl" → R-51 disclaimer template ZORUNLU (per-FAQ veya FAQ section başında); counter-argument FAQ aday (R-50).
• profile == "e-commerce" → product FAQ taxonomy (return/shipping/ sizing).
• profile == "b2b-saas" → technical FAQ taxonomy (integration/ pricing/security).
• profile == "local-service" → location-specific FAQ (service area/hours/contact, NAP signal).
• profile == "portfolio" → minimal FAQ 3-5 soft cap (R-09 10 hard cap profile-driven downgrade).

Step 3: R-09 FAQ Count Cap Enforce (Talep-Güdümlü)

• mode == "enhance": existing FAQ count read (Step 1 parse).
• mode == "add": target FAQ count plan (skill internal compose).
• R-09 talep-güdümlü band apply:
  • Kanıt varsa (PAA varlığı — serp_organic_live_advanced optional MCP; gerçek kullanıcı soruları — gsc__search_analytics query read) → 3-6 FAQ hedef.
  • Kanıt yoksa → daha az FAQ veya hiç (sıfır kabul; mekanik FAQ şişirme yasak — Principle 3).
  • Hard cap 10 (word count'tan bağımsız).
  • profile == "portfolio" → 5 soft cap (downgrade).

DURUR #2 trigger: mode == "add" AND faq_count > 10 → REJECT (R-09 violation, Principle 3 anti-cheap-content enforce; AI suistimal önlemi).

Step 4: R-43 Statik HTML Render (faq-block.template.html)

templates/content/faq-block.template.html render:

• Render-time slot:
  • {{QUESTION_N}} → FAQ question text (≤100 char, R-108 truncation; no : veya - R-04).
  • {{ANSWER_N}} → FAQ answer text (40-50 word definition snippet, R-107).
  • {primary_keyword} → new_content_plan.primary_keyword (heading keyword density %40-60 enforce).
  • {tone} → brand_identity.tone (profile-aware tone preserve).
  • {hitap} → brand_identity.hitap.
  • {primary_color} → brand_identity.primary_color (CSS scope).
  • {disclaimer} → content_settings.disclaimer_templates[profile] (profile==ymyl → ZORUNLU).

R-43 statik enforcement (CRITICAL — accordion YASAK):

• <details> element YASAK (rendered HTML grep RED).
• <summary> element YASAK (rendered HTML grep RED).
• display:none CSS YASAK (FAQ visible by default).
• JS accordion YASAK (toggleable interactive widget yasak).

DURUR #3 trigger: rendered HTML'de <details> veya <summary> detect → forbidden token grep RED, FAQ section discard, manuel revize gerekir.

Step 5: R-29 Pasaj Alıntılanabilirlik Gate

Her FAQ Q&A için pasaj alıntılanabilirlik test:

• H3 question (cevap-önce structure): "Soru?" formatında, soru cümlesi (no : veya - R-04).
• First sentence (TL;DR): cevap-önce 50-150 word, izole parse self-contained (H1/article context referansı YOK).
• Body (detail expansion): 50-300 word, source/citation embed.
• Total per FAQ: 100-450 word (R-29 + R-107 birleşik gate).

Pasaj alıntılanabilirlik test (LLM context window'a izole parse simulation):

• Self-contained check: "yukarıda gördüğünüz gibi" / "bu yazıda" / "önceki bölümde" referans regex → AMBER (R-01 + R-29 reuse).
• Cevap-önce mini paragraf intact: first sentence direct answer (interrogative reflection veya delay yasak).

Step 6: R-79 FAQPage @graph Schema Refresh + AIO Citation Pattern

JSON-LD @graph FAQPage entity update (R-79 + R-83 microdata yasak):

• mainEntity array: her FAQ → Question entity.
• Question.name: H3 question text (TL;DR truncate).
• Question.acceptedAnswer: Answer entity.
• Answer.text: full Q&A body (citation reference intact).

Schema validate (Google Rich Results Test API mock):

• @type: FAQPage present.
• mainEntity array length == FAQ count (R-09 cap compliance).
• Each Question.name non-empty.
• Each acceptedAnswer.text non-empty.

DURUR #4 trigger: Google Rich Results Test API mock validate REJECT → schema fix önce (FAQPage @graph entity broken), FAQ section yayın blocked.

AIO citation pattern enforcement (R-109/R-110/R-111):

• R-109 AIO Pattern: cevap-önce (R-29 reuse) + per 500 word min 1 max 2 citation (entity reference dense). FAQ acceptedAnswer.text içinde citation source URL bağlı (Principle 1 Layer 3 enforce).
• R-110 AIO Anti-Pattern: keyword stuffing FAQ Q yasak (heading keyword density >%60 AMBER); over-citation (>2 per 500w) yasak; fluff intro yasak (cevap-önce TL;DR direct).
• R-111 AIO Hijack (Quality-Driven): quality-driven enforce — R-44 source verification 3-katman intact, R-105 expert quote bank-driven; AI cliché phrase strip (tone_phrases_blocklist consume).

DURUR #6 trigger: AIO citation density per 500 word > 2 (over- citation Anti-Pattern R-110) → AMBER auto-correct (excess citation strip), retry once; persistent fail → RED.

Step 7: Foundational Principles 3-Katman Gate + events.jsonl Append

P1 truth-verifiable Layer 2 fact-check pass:

• Her FAQ Q&A için kaynak verify (R-44 source verification 3- katman: source URL reachable + content match + date freshness).
• Expert quote ise R-105 bank-driven verify (expert quote database lookup, no-match → reject).
• DURUR #5 trigger: P1 fact-check fail (FAQ Q&A uydurma claim/ source) → RED, FAQ section discard, dateModified bump YOK.

P2 profile-aware enforcement:

• profile == "ymyl" → R-51 disclaimer present check (per-FAQ veya section-level).
• counter-argument FAQ aday (R-50 cross-link, profile==ymyl).
• profile==e-commerce → product FAQ taxonomy verify.
• profile==b2b-saas → technical FAQ taxonomy verify.
• profile==local-service → NAP signal preserve.
• profile==portfolio → 5 soft cap enforce.

P3 anti-cheap-content enforcement:

• AI signature humanize (tone_phrases_blocklist consume; AI cliché regex strip).
• Heading keyword density per FAQ %40-60 enforce (primary_keyword vs FAQ Q heading; over-stuff >%60 AMBER, under <%40 AMBER).
• R-09 cap final check (talep-güdümlü 3-6 / hard cap 10 / 5 portfolio soft).

events.jsonl append (events.schema enum compliance):

• event_kind = work (ADR-020 production output enum).
• event_type = content_revise (events.schema 12-enum: FAQ enhance/add içerik revize semantik kanonik mapping).
• schema_version = 1.0 (events.schema const).
• event_id = UUID v4 (events.schema required).
• timestamp = UTC ISO 8601 (events.schema required).
• project_id = {input.project_slug} (events.schema pattern ^[a-z][a-z0-9-]*$ slug regex).
• target = {input.url}.
• mode = {input.mode} (enhance | add).
• faq_count_pre, faq_count_post (R-09 audit trail).
• aio_citation_density (R-109/R-110 audit trail).
• profile = project-config[profile] (P2 audit trail).

DURUR Conditions (6 koşul)

1. DURUR #1 — Existing Blog Yok (mode=enhance). outputs/blog/{slug}/article.html filesystem'de bulunamıyor → FAQ section parse edilemez (mode=enhance kaynak yok); kullanıcıyı new-blog skill'e yönlendir.
2. DURUR #2 — R-09 Cap Aşımı (mode=add). mode == "add" AND faq_count > 10 → REJECT (R-09 violation; AI suistimal Principle 3 anti-cheap-content enforce).
3. DURUR #3 — R-43 Accordion Detect. Rendered HTML'de <details> veya <summary> element detect → forbidden token grep RED, FAQ section discard, statik visible h3+p zorla.
4. DURUR #4 — R-79 FAQPage Schema Validate Fail. Google Rich Results Test API mock validate REJECT (mainEntity boş veya Question.name/acceptedAnswer.text broken) → schema fix önce, FAQ yayın blocked.
5. DURUR #5 — P1 Fact-Check Fail. Principle 1 Layer 2 post- generate fact-check fail (FAQ Q&A uydurma claim/source; R-44 source verify fail veya R-105 expert quote bank-driven match yok) → RED, FAQ section discard, dateModified bump YOK.
6. DURUR #6 — AIO Over-Citation (R-110 Anti-Pattern). AIO citation density per 500 word > 2 (over-citation) → AMBER auto- correct (excess citation strip), retry once; persistent fail → RED.

READ-ONLY Contract (F-6 Enforcement)

Bu skill master.xlsx hiçbir sheet'e YAZMAZ:

• master.xlsx[new_content_plan] allowed_writers null (F-6) → READ-ONLY consume (target URL row read, lifecycle_status check).
• master.xlsx[content_decay] allowed_writers null (F-6) → READ- ONLY consume (action trigger detect, mode=enhance trigger için existing FAQ detect).
• master.xlsx[content_improve] allowed_writers null (F-6, F-15) → READ-ONLY consume (FAQ optimization trigger detect; W-F3+W-F4 candidate input source).

transaction.append/update/delete call-site YASAK (Wave 1 paterni reuse: revise-content + new-blog precedent).

Sadece şu artifact'lere yazılır (outputs § + produces §):

• outputs/blog/{slug}/article.html (updated FAQ section + JSON-LD @graph FAQPage entity refresh).
• _state/events.jsonl (append, events.schema enum compliance).

WCAG 2.1 AA Accessibility (R-39)

FAQ block render:

• Heading hierarchy: h2 (FAQ section title) → h3 (each FAQ question); skip-level yasak.
• Focus-visible state: keyboard navigation FAQ list scroll (focus-visible CSS).
• Color contrast: primary_color vs background min 4.5:1 (WCAG 2.1 AA Level AA threshold).
• Static visible (R-43 accordion YASAK alignment): screen reader full FAQ content read (no toggleable hide).

Plugin-Agnostik Disiplin

Skill content'inde proje slug hardcode YASAK. Tüm proje referansları runtime input ({input.project_slug}, {input.url}) üzerinden çözümlenir. URL örnekleri SKILL.md'de generic placeholder (https://example.com/blog/post-slug) kullanır. Plugin agnostik MCP boundary korunur (.mcp.json hardcode slug yasak; F-16 Süleyman Seçenek D Wave 1 baseline reuse).

Versioning + Status

• version: "1.0" (Phase 11 Wave 2 ilk shipping).
• status: wip (Phase 11 Wave 3+ stabilizasyon sonrası active promote).
• Output schema_version: 1.0 (events.jsonl payload kontratı, events.schema const).