sergstack · sergstack · Aug 28, 2026 · Aug 28, 2026
@@ -25,11 +25,11 @@
 /PROMPT_QA_FACTORY.md @sergstack
 /PROJECT_CAPABILITIES.yaml @sergstack
 /knowledge_bundle_manifest.json @sergstack
-/CHATGPT_PROJECT_SYNC_CHECKLIST.md @sergstack
-/SMOKE_QA_RESULTS.md @sergstack
-/CROSS_PROJECT_SMOKE_QA_RESULTS.md @sergstack
-/PILOT_CASES.md @sergstack
-/PILOT_RESULTS_TEMPLATE.md @sergstack
+/docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md @sergstack
+/docs/evidence/SMOKE_QA_RESULTS.md @sergstack
+/docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md @sergstack
+/docs/operations/PILOT_CASES.md @sergstack
+/docs/operations/PILOT_RESULTS_TEMPLATE.md @sergstack
 /docs/AI_DEVELOPMENT_WORKFLOW.md @sergstack
 /docs/MERGE_GATE_OWNER_CHECKLIST.md @sergstack
 /schemas/ @sergstack

@@ -12,7 +12,7 @@
 - runtime_smoke_status: candidate
 - realistic_pilot_status: candidate (AI OS, Thinking, Analytics, and one cross-project routing/resume pilot passed; broader pilot set not run)
 - acceptance_status: candidate / ready for human review
-- smoke_qa_evidence: SMOKE_QA_RESULTS.md; CROSS_PROJECT_SMOKE_QA_RESULTS.md
+- smoke_qa_evidence: docs/evidence/SMOKE_QA_RESULTS.md; docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md
 - validation_gates: see `MASTER_STATUS.md` — "Validation Gates" and "Operational Gates" (canonical lists; do not copy them here)
 - blocked_items:
   - embeddings
@@ -51,7 +51,7 @@ Evidence-dependent or external state:
 Observed external pilot evidence — 2026-08-27:
 
 - `PILOT-AIOS-001` completed one live `[AI OS]` response and is recorded as
-  `candidate` with `medium` confidence in `PILOT_RESULTS_2026-08-27_AIOS.md`.
+  `candidate` with `medium` confidence in `docs/evidence/PILOT_RESULTS_2026-08-27_AIOS.md`.
 - The response named its KB sources, separated facts from hypotheses, retained
   the promotion gate for embeddings, semantic search, and vector DB, and
   routed the next step to bounded governance evidence collection.
@@ -61,7 +61,7 @@ Observed external pilot evidence — 2026-08-27:
 
 - `PILOT-THINKING-001` completed one live decision memo and is recorded as
   `candidate` with `medium` confidence in
-  `PILOT_RESULTS_2026-08-27_THINKING.md`.
+  `docs/evidence/PILOT_RESULTS_2026-08-27_THINKING.md`.
 - The memo compared four reversible options, separated facts, assumptions, and
   unknowns, identified risks, set a `recommended` decision status and revisit
   triggers, and handed the next stage back to `[AI OS]`.
@@ -71,7 +71,7 @@ Observed external pilot evidence — 2026-08-27:
 
 - `PILOT-ANALYTICS-001` completed one live quick-analysis response on an
   artificial three-row dataset and is recorded as `candidate` with `medium`
-  confidence in `PILOT_RESULTS_2026-08-27_ANALYTICS.md`.
+  confidence in `docs/evidence/PILOT_RESULTS_2026-08-27_ANALYTICS.md`.
 - The response defined grain, period, units, formulas, `RAW → stage → mart`,
   reconciliation checks, and limitations. It handled the zero-plan row without
   inventing a percentage and made no causal claim beyond the supplied data.
@@ -80,7 +80,7 @@ Observed external pilot evidence — 2026-08-27:
 
 - `PILOT-CROSS-001` completed one live `[AI OS] → [Thinking] → [AI OS]`
   routing/resume case and is recorded as `candidate` with `medium` confidence
-  in `PILOT_RESULTS_2026-08-27_CROSS.md`.
+  in `docs/evidence/PILOT_RESULTS_2026-08-27_CROSS.md`.
 - The route preserved the original goal, constraints, owner boundaries, and
   return path; no scope drift or role confusion was observed. This is limited
   evidence from one route, not proof of general cross-project reliability,
@@ -107,7 +107,8 @@ Recent verified state:
 - Goal Mode is the default; strict task packages are reserved for high-risk,
   already-scoped, ultra-long, or explicitly requested work.
 - ChatGPT Project upload mode is compact `Knowledge_Bundles` by default.
-- `SMOKE_QA_RESULTS.md` and `CROSS_PROJECT_SMOKE_QA_RESULTS.md` record
+- `docs/evidence/SMOKE_QA_RESULTS.md` and
+  `docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md` record
   2026-07-06 smoke QA evidence. Smoke QA does not equal production readiness.
 - `PILOT-AIOS-001`, `PILOT-THINKING-001`, `PILOT-ANALYTICS-001`, and
   `PILOT-CROSS-001` have recorded candidate results; all other pilots remain
@@ -162,7 +163,7 @@ Run repository validation before PR review: use the canonical command set from `
 
 Then complete operational verification:
 
-- `CHATGPT_PROJECT_SYNC_CHECKLIST.md`
+- `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`
 - obtain owner review for the four candidate pilots, then capture the next real
-  failure-to-regression case from `PILOT_CASES.md`
+  failure-to-regression case from `docs/operations/PILOT_CASES.md`
 - keep production promotion blocked until accepted pilot evidence exists
@@ -48,7 +48,7 @@
 Для поведения проекта также проверяй:
 `PROJECT_ROUTING.md`, `GOVERNANCE_RULES.md`, `AI_OS_WORKFLOW.md`, `HANDOFF_PROTOCOL.md`, `SMOKE_QA_FOR_AI_OS.md`, `ANTI_PATTERNS.md`.
 
-Если вопрос касается текущего статуса настройки проекта, сначала проверь `CURRENT_STATUS.md` и `SMOKE_QA_RESULTS.md`, если они доступны. Если их нет, прямо скажи, что текущий status setup не полностью записан.
+Если вопрос касается текущего статуса настройки проекта, сначала проверь `CURRENT_STATUS.md` и `docs/evidence/SMOKE_QA_RESULTS.md`, если они доступны. Если их нет, прямо скажи, что текущий status setup не полностью записан.
 
 ## 3. Обязательная шапка ответа
 

@@ -40,10 +40,10 @@ Canonical list. Other documents (including `CURRENT_STATUS.md`) must reference t
 
 ## Evidence Pointers
 
-- `SMOKE_QA_RESULTS.md` — `[AI OS]` smoke QA evidence from 2026-07-06.
-- `CROSS_PROJECT_SMOKE_QA_RESULTS.md` — cross-project smoke QA evidence from 2026-07-06.
-- `CHATGPT_PROJECT_SYNC_CHECKLIST.md` — manual sync status.
-- `PILOT_CASES.md` — pilot backlog; pilot completion still requires result evidence.
+- `docs/evidence/SMOKE_QA_RESULTS.md` — `[AI OS]` smoke QA evidence from 2026-07-06.
+- `docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md` — cross-project smoke QA evidence from 2026-07-06.
+- `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md` — manual sync status.
+- `docs/operations/PILOT_CASES.md` — pilot backlog; pilot completion still requires result evidence.
 - `StreamDeck/README.md` — StreamDeck active/candidate status.
 
 ## Repository State Reconciliation

@@ -48,9 +48,9 @@ production authorization.
 
 ## Operational Artifacts
 
-- `CHATGPT_PROJECT_SYNC_CHECKLIST.md`
-- `PILOT_CASES.md`
-- `SMOKE_QA_REFRESH_PLAN.md`
+- `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`
+- `docs/operations/PILOT_CASES.md`
+- `docs/operations/SMOKE_QA_REFRESH_PLAN.md`
 - `ChatGPT/[Project]/Knowledge_Bundles/`
 
 ## Canonical Workflows

@@ -33,11 +33,11 @@
 
 ## Operational Verification
 
-- `CHATGPT_PROJECT_SYNC_CHECKLIST.md`
-- `PROJECT_SYNC_TEMPLATE.md`
-- `PILOT_CASES.md`
-- `PILOT_RESULTS_TEMPLATE.md`
-- `SMOKE_QA_REFRESH_PLAN.md`
+- `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`
+- `docs/operations/PROJECT_SYNC_TEMPLATE.md`
+- `docs/operations/PILOT_CASES.md`
+- `docs/operations/PILOT_RESULTS_TEMPLATE.md`
+- `docs/operations/SMOKE_QA_REFRESH_PLAN.md`
 
 ## Knowledge Bundle Paths
 

@@ -83,7 +83,7 @@ SMOKE_QA_FOR_AI_OS.md
 
 ## 6. Operational verification
 
-After manual sync, record status in `CHATGPT_PROJECT_SYNC_CHECKLIST.md`, run smoke QA from `SMOKE_QA_REFRESH_PLAN.md`, and complete the relevant pilot in `PILOT_CASES.md`.
+After manual sync, record status in `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`, run smoke QA from `docs/operations/SMOKE_QA_REFRESH_PLAN.md`, and complete the relevant pilot in `docs/operations/PILOT_CASES.md`.
 
 ## 7. Compact Knowledge bundles
 
@@ -103,7 +103,7 @@ For a changed bundle, replace only that bundle in the matching ChatGPT Project:
    filename.
 
 Record the observed replacement and any affected smoke QA in
-`CHATGPT_PROJECT_SYNC_CHECKLIST.md`. A replaced filename proves only source
+`docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`. A replaced filename proves only source
 transport; it does not prove behavioural smoke QA, owner acceptance, or
 production authorization.
 

diff --git a/benchmarks/supermanager/COVERAGE.md b/benchmarks/supermanager/COVERAGE.md
@@ -24,4 +24,4 @@ All listed cases are deterministic Level A assertions. Model-evaluated and isola
 | Route: Codex APP | 3 | 1 | 0 | 0 | 4 | 0 |
 | Repository-wide hard-fail classes | 0 | 0 | 0 | 12 | 12 | 0 |
 
-The 22 regression cases preserve every individual `pass` row documented in `SMOKE_QA_RESULTS.md` and `CROSS_PROJECT_SMOKE_QA_RESULTS.md`; the cross-project aggregate `[AI OS]` row is represented by its seven underlying cases rather than counted twice.
+The 22 regression cases preserve every individual `pass` row documented in `docs/evidence/SMOKE_QA_RESULTS.md` and `docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md`; the cross-project aggregate `[AI OS]` row is represented by its seven underlying cases rather than counted twice.
diff --git a/benchmarks/supermanager/benchmark_definition.json b/benchmarks/supermanager/benchmark_definition.json
@@ -58,27 +58,27 @@
     {"id": "codex_app", "input_type": "Long-running Codex execution", "destination": "Codex APP directory", "positive_terms": ["long-running", "execution"], "negative_destination": "[Codex]"}
   ],
   "documented_smoke_cases": [
-    {"id": "aios_navigation", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 1 | Navigation |", "| pass |"]},
-    {"id": "aios_scope_routing", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 2 | Scope / Routing |", "| pass |"]},
-    {"id": "aios_evidence", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 3 | Evidence |", "| pass |"]},
-    {"id": "aios_governance", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 4 | Governance gates |", "| pass |"]},
-    {"id": "aios_handoff", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 5 | Handoff to Codex |", "| pass |"]},
-    {"id": "aios_goal_mode", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 6 | Goal Mode behavior |", "| pass |"]},
-    {"id": "aios_supervised_loop", "file": "SMOKE_QA_RESULTS.md", "tokens": ["| 7 | Supervised loop boundary |", "| pass |"]},
-    {"id": "thinking_decision", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | Decision memo |", "| pass |"]},
-    {"id": "thinking_judge", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | `@judge` review |", "| pass |"]},
-    {"id": "thinking_boundary", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | Routing boundary |", "| pass |"]},
-    {"id": "analytics_contract", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Data contract / stage / mart / QA / limitations |", "| pass |"]},
-    {"id": "analytics_codex_boundary", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Why not straight to Codex |", "| pass |"]},
-    {"id": "analytics_financial_boundary", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Financial task boundary |", "| pass |"]},
-    {"id": "llm_registry", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Prompt registry item |", "| pass |"]},
-    {"id": "llm_judge_revise", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Judge/revise unsupported claims |", "| pass |"]},
-    {"id": "llm_context_pack", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Context Pack |", "| pass |"]},
-    {"id": "codex_task_gate", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Task package gate |", "| pass |"]},
-    {"id": "codex_secrets", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Secrets / production deploy |", "| pass |"]},
-    {"id": "codex_goal_mode", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Goal Mode |", "| pass |"]},
-    {"id": "router_classification", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | Classify 10 raw inputs |", "| pass |"]},
-    {"id": "router_multi_project", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | Multi-project input |", "| pass |"]},
-    {"id": "router_streamdeck", "file": "CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | StreamDeck autoloop + PR route |", "| pass |"]}
+    {"id": "aios_navigation", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 1 | Navigation |", "| pass |"]},
+    {"id": "aios_scope_routing", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 2 | Scope / Routing |", "| pass |"]},
+    {"id": "aios_evidence", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 3 | Evidence |", "| pass |"]},
+    {"id": "aios_governance", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 4 | Governance gates |", "| pass |"]},
+    {"id": "aios_handoff", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 5 | Handoff to Codex |", "| pass |"]},
+    {"id": "aios_goal_mode", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 6 | Goal Mode behavior |", "| pass |"]},
+    {"id": "aios_supervised_loop", "file": "docs/evidence/SMOKE_QA_RESULTS.md", "tokens": ["| 7 | Supervised loop boundary |", "| pass |"]},
+    {"id": "thinking_decision", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | Decision memo |", "| pass |"]},
+    {"id": "thinking_judge", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | `@judge` review |", "| pass |"]},
+    {"id": "thinking_boundary", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Thinking]` | Routing boundary |", "| pass |"]},
+    {"id": "analytics_contract", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Data contract / stage / mart / QA / limitations |", "| pass |"]},
+    {"id": "analytics_codex_boundary", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Why not straight to Codex |", "| pass |"]},
+    {"id": "analytics_financial_boundary", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Analytics]` | Financial task boundary |", "| pass |"]},
+    {"id": "llm_registry", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Prompt registry item |", "| pass |"]},
+    {"id": "llm_judge_revise", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Judge/revise unsupported claims |", "| pass |"]},
+    {"id": "llm_context_pack", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[LLM]` | Context Pack |", "| pass |"]},
+    {"id": "codex_task_gate", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Task package gate |", "| pass |"]},
+    {"id": "codex_secrets", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Secrets / production deploy |", "| pass |"]},
+    {"id": "codex_goal_mode", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Codex]` | Goal Mode |", "| pass |"]},
+    {"id": "router_classification", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | Classify 10 raw inputs |", "| pass |"]},
+    {"id": "router_multi_project", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | Multi-project input |", "| pass |"]},
+    {"id": "router_streamdeck", "file": "docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md", "tokens": ["| `[Inbox Router]` | StreamDeck autoloop + PR route |", "| pass |"]}
   ]
 }
diff --git a/docs/REPOSITORY_MAP.md b/docs/REPOSITORY_MAP.md
@@ -72,9 +72,9 @@ debugging a sync problem.
 | Entry and routing | `README.md`, this map, registries | Explain where to start; do not copy domain methodology |
 | Governance and contracts | `AGENTS.md`, `GOAL_MODE.md`, active standards | Normative rules; protected-path review applies |
 | Current status | root and project-local `CURRENT_STATUS.md` | State maturity and next action; do not redefine gates |
+| Operations | [`docs/operations/`](operations) | Active sync, smoke-QA, and pilot records; not entry-point policy |
 | Gate definitions | `MASTER_STATUS.md` | Define validation and operational gates and point to evidence |
-| Operational records | checklists, pilot plans, manifests | Track one operation or package; not general policy |
-| Evidence | smoke-QA, pilot, and acceptance results | Record what was checked; passing evidence is not production approval |
+| Evidence | [`docs/evidence/`](evidence) | Record what was checked; passing evidence is not production approval |
 | Delivery artifacts | `Knowledge_Bundles/`, Stream Deck exports | Distributable surfaces derived from canonical sources |
 | History | [`archive/`](../archive), including completed task packages | Preserve audit history; not current guidance unless actively referenced |
 

@@ -20,10 +20,10 @@ Revision update: the two original `revise` findings from PR #62 were fixed in Pr
 
 ## Files Checked
 
-- `CHATGPT_PROJECT_SYNC_CHECKLIST.md`
+- `../operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`
 - `SMOKE_QA_RESULTS.md`
-- `SMOKE_QA_REFRESH_PLAN.md`
-- `PILOT_CASES.md`
+- `../operations/SMOKE_QA_REFRESH_PLAN.md`
+- `../operations/PILOT_CASES.md`
 - `archive/reports/KB_COMPACT_CONSISTENCY_REPORT.md`
 - `ChatGPT/[AI OS]/PROJECT_INSTRUCTIONS.md`
 - `ChatGPT/[Thinking]/PROJECT_INSTRUCTIONS.md`
@@ -83,7 +83,7 @@ Other project smoke QA was run in ChatGPT UI as one combined prompt per project,
 | `python3 scripts/check_knowledge_bundles.py` | pass | 6 projects checked; 30 bundles checked; failed=0. |
 | `python3 scripts/sync_aios.py` | pass | Internal checks passed; helper printed sync readiness guidance and did not upload, push, or modify remote systems. |
 | `git diff --check` | pass | No whitespace errors reported. |
-| `git status --short` | pass | Shows only expected docs/settings files changed: `CHATGPT_PROJECT_SYNC_CHECKLIST.md`, `CROSS_PROJECT_SMOKE_QA_RESULTS.md`, `ChatGPT/[Codex]/PROJECT_INSTRUCTIONS.md`, and `ChatGPT/[Inbox Router]/PROJECT_INSTRUCTIONS.md`. |
+| `git status --short` | pass | Shows only expected docs/settings files changed: `docs/operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`, `docs/evidence/CROSS_PROJECT_SMOKE_QA_RESULTS.md`, `ChatGPT/[Codex]/PROJECT_INSTRUCTIONS.md`, and `ChatGPT/[Inbox Router]/PROJECT_INSTRUCTIONS.md`. |
 
 ## Checklist Updates Made
 
@@ -97,7 +97,7 @@ Other project smoke QA was run in ChatGPT UI as one combined prompt per project,
 
 ## Pilot Statuses
 
-`PILOT_CASES.md` was not changed.
+`../operations/PILOT_CASES.md` was not changed.
 All pilot completions remain open / `not_verified` unless already recorded elsewhere.
 
 ## Residual Risks
@@ -114,7 +114,7 @@ All pilot completions remain open / `not_verified` unless already recorded elsew
 - Checklist updates: observed UI sync and smoke QA status recorded for `[Thinking]`, `[Analytics]`, `[LLM]`, `[Codex]`, and `[Inbox Router]`; `[AI OS]` reused issue #59 row.
 - Pilots left open: yes.
 - Risks / limitations: combined prompts for the original cross-project run, no full byte-for-byte UI instruction comparison, no production promotion.
-- Rollback: revert this documentation update, the related `CHATGPT_PROJECT_SYNC_CHECKLIST.md` row updates, and the two Project Instructions wording updates.
+- Rollback: revert this documentation update, the related `../operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md` row updates, and the two Project Instructions wording updates.
 - Acceptance status: `candidate / ready for human review`.
 
 ## Next Step

diff --git a/PILOT_RESULTS_2026-08-27_AIOS.md → ...evidence/PILOT_RESULTS_2026-08-27_AIOS.md b/PILOT_RESULTS_2026-08-27_AIOS.md → ...evidence/PILOT_RESULTS_2026-08-27_AIOS.md
@@ -34,7 +34,7 @@ Evidence:
 - direct live response: <https://chatgpt.com/g/g-p-6a0512a228c88191afcc953866789dad-ai-os/c/6a9019be-c360-83eb-b5fb-0d2bd736cd7a>;
 - response completed in the live `[AI OS]` project; no source upload,
   settings change, implementation handoff, or production action occurred;
-- repository context: `PILOT_CASES.md`, `CHATGPT_PROJECT_SYNC_CHECKLIST.md`,
+- repository context: `../operations/PILOT_CASES.md`, `../operations/CHATGPT_PROJECT_SYNC_CHECKLIST.md`,
   and the existing `[AI OS]` smoke/sync evidence.
 
 Checks run:

diff --git a/PILOT_RESULTS_2026-08-27_ANALYTICS.md → ...nce/PILOT_RESULTS_2026-08-27_ANALYTICS.md b/PILOT_RESULTS_2026-08-27_ANALYTICS.md → ...nce/PILOT_RESULTS_2026-08-27_ANALYTICS.md
diff --git a/PILOT_RESULTS_2026-08-27_CROSS.md → ...vidence/PILOT_RESULTS_2026-08-27_CROSS.md b/PILOT_RESULTS_2026-08-27_CROSS.md → ...vidence/PILOT_RESULTS_2026-08-27_CROSS.md