fix: align codex skill runtime docs

Author: boshu2

skills-codex/.agentops-manifest.json

@@ -658,7 +658,7 @@
       "name": "post-mortem",
       "source_skill": "skills/post-mortem",
       "source_hash": "fbbedf5eb3b7883680a8a538748816756c7daa1d6d80b450dc2830af3fa9da17",
-      "generated_hash": "3d1b1850b2d8d15ecc129f7438a1d809bb97c989e9d47da1d84926c2515f40c1"
+      "generated_hash": "530bb54713d3b06576418e4c74630e71e886b97299ffa9d10d0a0247cb50647d"
     },
     {
       "name": "pr-implement",
@@ -826,7 +826,7 @@
       "name": "using-agentops",
       "source_skill": "skills/using-agentops",
       "source_hash": "4416cf30a11c6486e3962e9ed2e44067fab2b17bcc498b759d15beb6f1ae504c",
-      "generated_hash": "ac7744194ca5f75d2b34bc83bc9e5dec25b5c64893b74e9501724a1134616919"
+      "generated_hash": "a3466b4aaa05e2c126783bc26156d7fdaf0876a6493cad2c9dd058f770da1c21"
     },
     {
       "name": "validation",

skills-codex/post-mortem/.agentops-generated.json

@@ -3,5 +3,5 @@
   "source_skill": "skills/post-mortem",
   "layout": "modular",
   "source_hash": "fbbedf5eb3b7883680a8a538748816756c7daa1d6d80b450dc2830af3fa9da17",
-  "generated_hash": "3d1b1850b2d8d15ecc129f7438a1d809bb97c989e9d47da1d84926c2515f40c1"
+  "generated_hash": "530bb54713d3b06576418e4c74630e71e886b97299ffa9d10d0a0247cb50647d"
 }

skills-codex/post-mortem/SKILL.md

@@ -30,8 +30,6 @@ $post-mortem --process-only     # skip council+extraction, run Phase 3-5 on back
 $post-mortem --skip-activate    # extract + process but don't write MEMORY.md
 $post-mortem --deep recent      # thorough council review
 $post-mortem --mixed epic-123   # cross-vendor (Claude + Codex)
-$post-mortem --explorers=2 epic-123  # deep investigation before judging
-$post-mortem --debate epic-123      # two-round adversarial review
 $post-mortem --skip-checkpoint-policy epic-123  # skip ratchet chain validation
 ```
 
@@ -44,7 +42,7 @@ ao codex stop
 ao codex status
 ```
 
-`ao codex stop` resolves the latest Codex transcript or history fallback, queues/persists learnings safely, and runs close-loop maintenance without depending on runtime hooks.
+`ao codex stop` uses the latest transcript or history fallback to queue/persist learnings and run close-loop maintenance without runtime hooks.
 
 ---
 
@@ -988,40 +986,28 @@ $rpi "<highest-priority enhancement>"
 **User says:** `$post-mortem`
 
 **What happens:**
-1. Agent scans recent commits (last 7 days)
-2. Runs `$council --deep --preset=retrospective validate recent`
-3. 3 judges (plan-compliance, tech-debt, learnings) review
-4. Extracts learnings inline (Phase 2: context gathering, classification, writing)
-5. Processes backlog (Phase 3: scores, deduplicates, flags stale)
-6. Activates high-value learnings (Phase 4: promotes to MEMORY.md)
-7. Retires stale knowledge (Phase 5)
-8. Synthesizes process improvement proposals
-9. Harvests next-work items to `.agents/rpi/next-work.jsonl`
-10. Feeds learnings to knowledge flywheel via `ao forge`
-
-**Result:** Post-mortem report with learnings, tech debt identified, knowledge lifecycle stats, and suggested next `$rpi` command.
+1. Agent scans recent commits.
+2. Runs `$council --deep validate recent`.
+3. Extracts learnings, processes backlog, and promotes items.
+4. Harvests next-work to `.agents/rpi/next-work.jsonl`.
 
-### Other Modes
+**Result:** Report with learnings, stats, and a suggested `$rpi` command.
 
-- **Epic-specific:** `$post-mortem ag-5k2` — loads plan, council reviews vs plan, full Phase 2-6 lifecycle
-- **Quick capture:** `$post-mortem --quick "insight"` — writes directly to `.agents/learnings/`, no council
-- **Process-only:** `$post-mortem --process-only` — skip council/extraction, run Phase 3-5 on backlog
-- **Cross-vendor:** `$post-mortem --mixed ag-3b7` — 3 Claude + 3 Codex judges for broader coverage
+### Other Modes
 
-**Result:** Higher confidence validation with cross-vendor review before closing epic.
+- **Epic-specific:** `$post-mortem ag-5k2` — review against the target plan
+- **Quick capture:** `$post-mortem --quick "insight"` — write a learning without council
+- **Process-only:** `$post-mortem --process-only` — run backlog processing only
+- **Cross-vendor:** `$post-mortem --mixed ag-3b7` — broaden judgment coverage
 
 ## Troubleshooting
 
 | Problem | Cause | Solution |
 |---------|-------|----------|
 | Council times out | Epic too large or too many files changed | Split post-mortem into smaller reviews or increase timeout |
 | No next-work items harvested | Council found no tech debt or improvements | Flywheel stable — write entry with empty items array to next-work.jsonl |
-| Schema validation failed | Harvested item missing required field or has invalid enum value | Drop invalid item, log error, proceed with valid items only |
 | Checkpoint-policy preflight blocks | Prior FAIL verdict in ratchet chain without fix | Resolve prior failure (fix + re-vibe) or skip checkpoint-policy via `--skip-checkpoint-policy` |
 | Metadata verification fails | Plan vs actual files mismatch or missing cross-references | Include failures in council packet as `context.metadata_failures` — judges assess severity |
-| Phase 3 finds zero learnings | last-processed marker is very recent or no learnings exist | Reset marker: `date -v-30d +%Y-%m-%dT%H:%M:%S > .agents/ao/last-processed` |
-| Phase 4 promotion duplicates | MEMORY.md already has the insight | Grep-based dedup should catch this; if not, manually deduplicate MEMORY.md |
-| Phase 5 archives too aggressively | 30-day window too short for slow-cadence projects | Adjust the staleness threshold in `references/backlog-processing.md` |
 
 ---
 
@@ -1074,4 +1060,3 @@ $rpi "<highest-priority enhancement>"
 - `scripts/preflight-refs.sh`
 - `scripts/validate.sh`
 - `scripts/write-evidence-only-closure.sh`
-

skills-codex/using-agentops/.agentops-generated.json

@@ -3,5 +3,5 @@
   "source_skill": "skills/using-agentops",
   "layout": "modular",
   "source_hash": "4416cf30a11c6486e3962e9ed2e44067fab2b17bcc498b759d15beb6f1ae504c",
-  "generated_hash": "ac7744194ca5f75d2b34bc83bc9e5dec25b5c64893b74e9501724a1134616919"
+  "generated_hash": "a3466b4aaa05e2c126783bc26156d7fdaf0876a6493cad2c9dd058f770da1c21"
 }

skills-codex/using-agentops/SKILL.md

@@ -168,9 +168,8 @@ AgentOps has three runtime modes. Do not assume hook automation exists everywher
 
 | Mode | When it applies | Start path | Closeout path | Guarantees |
 |------|-----------------|------------|---------------|------------|
-| `hook-capable` | Claude/OpenCode with lifecycle hooks installed | Runtime hook or `ao inject` / `ao lookup` | Runtime hook or `ao forge transcript` + `ao flywheel close-loop` | Automatic startup/context injection and session-end maintenance when hooks are installed |
 | `codex-hookless-fallback` | Codex Desktop / Codex CLI without hook surfaces | `ao codex start` | `ao codex stop` | Explicit startup context, citation tracking, transcript fallback, and close-loop metrics without hooks |
-| `manual` | No hooks and no Codex-native runtime detection | `ao inject` / `ao lookup` | `ao forge transcript` + `ao flywheel close-loop` | Works everywhere, but lifecycle actions are operator-driven |
+| `manual` | Codex cannot resolve repo/runtime state automatically | `ao inject` / `ao lookup` | `ao forge transcript` + `ao flywheel close-loop` | Works everywhere, but lifecycle actions are operator-driven |
 
 ## Issue Tracking
 
@@ -187,20 +186,13 @@ bd vc status          # Inspect Dolt state if needed (JSONL auto-sync is automat
 
 ### Startup Context Loading
 
-**Hook-capable runtimes**
-1. `session-start.sh` (or equivalent) can run at session start.
-2. In `manual` mode, MEMORY.md is auto-loaded and the hook points to on-demand retrieval (`ao search`, `ao lookup`).
-3. In `lean` mode, the hook extracts pending knowledge and injects prior learnings with a reduced token budget.
-4. This skill can be injected automatically into session context.
-
-**Codex hookless fallback**
-1. Run `ao codex start`.
+1. Run `ao codex start` when Codex is detected without hooks.
 2. AgentOps inspects `.agents/`, runs safe close-loop maintenance, syncs MEMORY.md, and writes `.agents/ao/codex/startup-context.md`.
 3. Surfaced learnings, patterns, and findings are cited as `retrieved`.
-4. Use `ao lookup` for automatic citations during work, or `ao search --cite retrieved|reference|applied` when a search result is actually adopted.
+4. Use `ao lookup` for automatic citations during work, or `ao search --cite retrieved|reference|applied` when a search result is adopted.
 5. End the session with `ao codex stop`, then verify loop health with `ao codex status`.
 
-**Result:** The agent gets the RPI workflow, prior context, and a citation path in both modes. The difference is whether lifecycle work is hook-driven or command-driven.
+**Result:** In hookless Codex mode, the agent still gets prior context, citations, and closeout without hidden hooks.
 
 ### Workflow Reference During Planning
 
@@ -229,4 +221,3 @@ bd vc status          # Inspect Dolt state if needed (JSONL auto-sync is automat
 ### scripts/
 
 - `scripts/validate.sh`
-

skills/post-mortem/SKILL.md

@@ -43,8 +43,6 @@ Six phases:
 /post-mortem --skip-activate    # extract + process but don't write MEMORY.md
 /post-mortem --deep recent      # thorough council review
 /post-mortem --mixed epic-123   # cross-vendor (Claude + Codex)
-/post-mortem --explorers=2 epic-123  # deep investigation before judging
-/post-mortem --debate epic-123      # two-round adversarial review
 /post-mortem --skip-checkpoint-policy epic-123  # skip ratchet chain validation
 ```
 
@@ -57,7 +55,7 @@ ao codex stop
 ao codex status
 ```
 
-`ao codex stop` resolves the latest Codex transcript or history fallback, queues/persists learnings safely, and runs close-loop maintenance without depending on runtime hooks.
+`ao codex stop` uses the latest transcript or history fallback to queue/persist learnings and run close-loop maintenance without runtime hooks.
 
 ---
 
@@ -980,40 +978,28 @@ Ship it
 **User says:** `/post-mortem`
 
 **What happens:**
-1. Agent scans recent commits (last 7 days)
-2. Runs `/council --deep --preset=retrospective validate recent`
-3. 3 judges (plan-compliance, tech-debt, learnings) review
-4. Extracts learnings inline (Phase 2: context gathering, classification, writing)
-5. Processes backlog (Phase 3: scores, deduplicates, flags stale)
-6. Activates high-value learnings (Phase 4: promotes to MEMORY.md)
-7. Retires stale knowledge (Phase 5)
-8. Synthesizes process improvement proposals
-9. Harvests next-work items to `.agents/rpi/next-work.jsonl`
-10. Feeds learnings to knowledge flywheel via `ao forge`
-
-**Result:** Post-mortem report with learnings, tech debt identified, knowledge lifecycle stats, and suggested next `/rpi` command.
+1. Agent scans recent commits.
+2. Runs `/council --deep validate recent`.
+3. Extracts learnings, processes backlog, and promotes items.
+4. Harvests next-work to `.agents/rpi/next-work.jsonl`.
 
-### Other Modes
+**Result:** Report with learnings, stats, and a suggested `/rpi` command.
 
-- **Epic-specific:** `/post-mortem ag-5k2` — loads plan, council reviews vs plan, full Phase 2-6 lifecycle
-- **Quick capture:** `/post-mortem --quick "insight"` — writes directly to `.agents/learnings/`, no council
-- **Process-only:** `/post-mortem --process-only` — skip council/extraction, run Phase 3-5 on backlog
-- **Cross-vendor:** `/post-mortem --mixed ag-3b7` — 3 Claude + 3 Codex judges for broader coverage
+### Other Modes
 
-**Result:** Higher confidence validation with cross-vendor review before closing epic.
+- **Epic-specific:** `/post-mortem ag-5k2` — review against the target plan
+- **Quick capture:** `/post-mortem --quick "insight"` — write a learning without council
+- **Process-only:** `/post-mortem --process-only` — run backlog processing only
+- **Cross-vendor:** `/post-mortem --mixed ag-3b7` — broaden judgment coverage
 
 ## Troubleshooting
 
 | Problem | Cause | Solution |
 |---------|-------|----------|
 | Council times out | Epic too large or too many files changed | Split post-mortem into smaller reviews or increase timeout |
 | No next-work items harvested | Council found no tech debt or improvements | Flywheel stable — write entry with empty items array to next-work.jsonl |
-| Schema validation failed | Harvested item missing required field or has invalid enum value | Drop invalid item, log error, proceed with valid items only |
 | Checkpoint-policy preflight blocks | Prior FAIL verdict in ratchet chain without fix | Resolve prior failure (fix + re-vibe) or skip checkpoint-policy via `--skip-checkpoint-policy` |
 | Metadata verification fails | Plan vs actual files mismatch or missing cross-references | Include failures in council packet as `context.metadata_failures` — judges assess severity |
-| Phase 3 finds zero learnings | last-processed marker is very recent or no learnings exist | Reset marker: `date -v-30d +%Y-%m-%dT%H:%M:%S > .agents/ao/last-processed` |
-| Phase 4 promotion duplicates | MEMORY.md already has the insight | Grep-based dedup should catch this; if not, manually deduplicate MEMORY.md |
-| Phase 5 archives too aggressively | 30-day window too short for slow-cadence projects | Adjust the staleness threshold in `references/backlog-processing.md` |
 
 ---