oh-my-customcodex 0.5.8 → 0.5.10

package/templates/.claude/skills/dev-lead-routing/SKILL.md

@@ -104,24 +104,17 @@ Check if Agent Teams is available (`OMCODEX_AGENT_TEAMS=1` or TeamCreate/SendMes
 | Cross-layer debugging (FE + BE + DB) | Agent Teams |
 | Simple file search/validation | Task Tool |
 
-### Step 2: Codex-Exec Hybrid (Implementation Tasks)
+### Step 2: External Interop Guidance (Implementation Tasks)
 For **new file creation**, **boilerplate**, or **test code generation**:
 
-1. Check `/tmp/.codex-env-status-*` for codex, gemini, and rtk availability
-2. If codex available AND task involves new file creation → automatically delegate to `/codex-exec` for scaffolding:
-   - Display: `[Codex Hybrid] Delegating to codex-exec...`
-   - codex-exec generates initial code (strength: fast generation)
-   - Selected Claude expert reviews and refines codex output (strength: reasoning, quality)
-3. If codex unavailable but gemini available → delegate to `/gemini-exec` for scaffolding:
-   - Display: `[Gemini Hybrid] Delegating to gemini-exec...`
-   - gemini-exec generates initial code
-   - Selected Claude expert reviews and refines output
-4. If RTK available (`RTK=available` in env status) → optionally wrap Claude expert output through RTK to reduce token consumption by 60-90%:
+1. Use the selected expert agent as the default implementation path.
+2. If the native Claude Code plugin `openai/codex-plugin-cc` is explicitly installed and requested, it may provide Codex interop for scaffolding before expert review.
+3. If RTK is available (`RTK=available` in env status), optionally wrap expert output through `rtk-exec` to reduce token consumption by 60-90%:
    - Display: `[RTK Proxy] Token optimization active via rtk-exec`
    - RTK acts as a transparent proxy — no change to expert selection
-5. If none available → display `[External CLI] Unavailable — proceeding with {expert} directly` and use Claude expert directly
+4. Otherwise display `[External CLI] Not requested — proceeding with {expert} directly` and use the expert directly.
 
-**Suitable**: New file creation, boilerplate, scaffolding, test code
+**Suitable for optional plugin interop**: New file creation, boilerplate, scaffolding, test code
 **Unsuitable**: Existing code modification, architecture decisions, bug fixes
 
 ### Step 3: Expert Agent Selection

package/templates/.claude/skills/intent-detection/SKILL.md

@@ -262,14 +262,12 @@ english:
 ```
 Research intent detected (confidence >= 70%)
     ↓
-Check Codex CLI availability
-    ├─ Available (codex binary + OPENAI_API_KEY)
-    │   → Use codex-exec skill with --effort xhigh
-    │   → Prompt: "Research and analyze: {user_request}"
-    │   → Returns: structured findings for orchestrator
-    └─ Unavailable
-        → Fall back to Claude's WebFetch/WebSearch
-        → Orchestrator handles directly or via general-purpose agent
+Use the research workflow or researcher/expert agents
+    ├─ Native Claude Code plugin `openai/codex-plugin-cc` explicitly installed/requested
+    │   → Optional Codex interop may assist analysis
+    └─ Otherwise
+        → Use existing expert agents, researcher, WebFetch/WebSearch, or RTK paths
+        → Orchestrator handles directly or via specialist agent
 ```
 
 ### Confidence Scoring
@@ -288,6 +286,6 @@ Check Codex CLI availability
 ├── Input: "{user input}"
 ├── Workflow: research-workflow
 ├── Confidence: {percentage}%
-├── Method: codex-exec (xhigh) | WebFetch fallback
+├── Method: research workflow | researcher/expert agent | optional Codex plugin when requested
 └── Reason: {explanation}
 ```

package/templates/.claude/skills/research/SKILL.md

@@ -167,17 +167,8 @@ Batch 3: T9, T10            (Innovation)
 
 Before starting verification rounds, check codex availability:
 
-```bash
-# Run this check once before Phase 2 begins
-which codex &>/dev/null && [ -n "$OPENAI_API_KEY" ]
-# Exit 0 → codex available: enable dual-model verification (opus + codex)
-# Exit 1 → codex unavailable: display notice and proceed with opus-only
-```
-
-If unavailable, display: `[Phase 2] Codex unavailable — opus-only verification`
-
 ```
-Team findings ──→ opus 4.6 verification ──→ codex-exec xhigh verification (if available)
+Team findings ──→ opus 4.6 verification ──→ optional plugin-assisted verification
        │                                              │
        └── Contradiction detected? ── YES ──→ Round N+1
                                       NO  ──→ Consensus reached → Phase 3
@@ -185,8 +176,7 @@ Team findings ──→ opus 4.6 verification ──→ codex-exec xhigh verific
 
 Each round:
 1. **opus 4.6**: Deep reasoning verification — checks logical consistency, identifies gaps, challenges assumptions
-2. **codex-exec xhigh** (when available): Independent code-level verification — validates technical claims, tests feasibility
-   - If unavailable: display `[Phase 2] Round {N}: Codex unavailable, proceeding with opus verification only`
+2. **Optional Codex interop**: Use native Claude Code plugin `openai/codex-plugin-cc` only when explicitly installed/requested; otherwise use researcher or RTK-backed local evidence gathering.
 3. **Contradiction resolution**: Reconcile divergent findings between teams and verifiers
 4. **Convergence check**: All major claims verified with no outstanding contradictions → proceed
 
@@ -319,16 +309,11 @@ Round N:
           - Internal consistency (breadth ↔ depth alignment)
           - Cross-domain consistency (security ↔ architecture)
           - Evidence quality (claims without backing)
-  Step 2: codex-exec validates technical claims (when available):
-          a. Invoke: /codex-exec with findings from all teams
-          b. Prompt:  "Validate technical claims: {findings}.
-                       Check code patterns, benchmark reproducibility,
-                       dependency resolution."
-          c. Effort:  --effort xhigh
-          d. Parse:   contradictions → merge with opus findings
-          e. On timeout/error: log "[Phase 2] Round {N}: codex-exec error — {reason},
-                                continuing with opus results only"
-     If unavailable: log "[Phase 2] Round {N}: Codex unavailable, proceeding with opus verification only"
+  Step 2: Optional plugin-assisted validation (only if `openai/codex-plugin-cc` is explicitly installed/requested):
+          a. Validate technical claims against code patterns, benchmark reproducibility,
+             and dependency resolution.
+          b. Parse contradictions → merge with opus findings.
+          c. On unavailable plugin or error: log "[Phase 2] Round {N}: optional Codex interop unavailable — continuing with opus/researcher results".
   Step 3: Compile contradiction list
           - 0 contradictions → CONVERGED
           - >0 contradictions → feedback to relevant teams → Round N+1
@@ -410,7 +395,7 @@ This advisory is informational only and does not block execution.
 
 | Scenario | Fallback |
 |----------|----------|
-| codex-exec unavailable | opus-only verification (still min 2 rounds) |
+| Optional Codex plugin unavailable | opus/researcher verification (still min 2 rounds) |
 | Agent Teams unavailable | Standard Agent tool with R009 batching |
 | Partial team failure | Synthesize from available results, note gaps in report |
 | GitHub issue creation fails | Output report to conversation only |

package/templates/.claude/skills/roundtable-debate/SKILL.md

@@ -12,7 +12,7 @@ version: 1.0.0
 
 ## Purpose
 
-Run a bounded debate when convergence would hide useful disagreement. Unlike `agora`, which drives toward consensus, this workflow preserves minority positions and requires explicit justification before dismissing them.
+Run a bounded debate when convergence would hide useful disagreement. Use this workflow when debate should preserve minority positions and require explicit justification before dismissing them.
 
 ## When To Use
 
@@ -54,11 +54,10 @@ Hard cap: two debate rounds. If the decision still depends on missing facts, sto
 {adopt | defer | reject | gather-more-evidence}
 ```
 
-## Relationship To Agora
+## Relationship To Other Review Workflows
 
 | Workflow | Goal | Best For |
 |----------|------|----------|
-| `agora` | adversarial consensus | release gates, spec approval |
 | `roundtable-debate` | dissent preservation | ambiguous strategy, architectural tradeoffs |
 
-Use `agora --anti-groupthink` when you need consensus plus explicit dissent handling.
+Use `roundtable-debate` when current guidance needs adversarial review plus explicit dissent handling.

package/templates/.claude/skills/skill-extractor/SKILL.md

@@ -1,150 +1,239 @@
 ---
 name: skill-extractor
-description: Analyze task trajectories to propose reusable SKILL.md candidates from successful patterns
+description: Analyze recurring task trajectories and evidence to propose reusable workflow packaging candidates
 scope: core
 user-invocable: true
-argument-hint: "[--threshold <n>] [--dry-run]"
-version: 1.0.0
+argument-hint: "[--threshold <n>] [--dry-run] [--all]"
+version: 1.1.0
 ---
 
 # Skill Extractor
 
-Analyze completed task outcomes to identify reusable patterns and propose new SKILL.md candidates. Inspired by Hermes Agent's self-learning skill extraction — adapted for oh-my-customcodex's compilation metaphor.
+Analyze completed task outcomes and recent work evidence to identify recurring workflows that may deserve reusable packaging. Inspired by Hermes Agent's self-learning skill extraction — adapted for oh-my-customcodex's compilation metaphor.
 
 ## Philosophy
 
-In the compilation metaphor: task trajectories are runtime traces, and extracted skills are new source code. This skill turns successful execution patterns into reusable knowledge artifacts.
+In the compilation metaphor: task trajectories are runtime traces, and extracted skills are new source code. This skill turns repeated, successful execution patterns into reusable knowledge artifacts only after evidence review and user approval.
 
 ```
-Runtime traces (task outcomes) → Pattern analysis → SKILL.md proposal → User approval → mgr-creator
+Runtime traces + memory + rollout summaries + inventory → Evidence-first shortlist → Packaging recommendation → User approval → mgr-creator or automation owner
 ```
 
 ## Usage
 
 ```
-/skill-extractor                    # Analyze current session outcomes
-/skill-extractor --threshold 2      # Lower success threshold (default: 3)
-/skill-extractor --dry-run          # Preview proposals without writing
+/skill-extractor                    # Analyze current session outcomes and local evidence
+/skill-extractor --threshold 2      # Lower recurring evidence threshold (default: 3)
+/skill-extractor --dry-run          # Preview shortlist without writing
+/skill-extractor --all              # Include broader session and memory history when available
 ```
 
 ## Options
 
 ```
---threshold, -t   Minimum success count for pattern qualification (default: 3)
---dry-run, -d     Preview proposals to stdout only, no file writes
---all             Include all sessions (not just current, requires task outcome history)
+--threshold, -t   Minimum evidence count for recurring-workflow qualification (default: 3)
+--dry-run, -d     Preview candidates to stdout only, no file writes
+--all             Include all sessions and available memory/history, not just current session
 ```
 
 ## Workflow
 
-### Phase 1: Collect Task Outcomes
+### Phase 1: Evidence-First Candidate Discovery
 
-Read task outcome data from the session:
+Collect candidates from concrete, dated evidence before proposing any packageable artifact. Prefer local evidence first; use optional integrations only when available.
 
-```bash
-# Current session outcomes (from task-outcome-recorder hook)
-OUTCOMES_FILE="/tmp/.codex-task-outcomes-${PPID}"
-```
+Required and recommended sources:
+
+1. **Recent session outcomes** from the task-outcome-recorder hook:
+
+   ```bash
+   # Current session outcomes
+   OUTCOMES_FILE="/tmp/.codex-task-outcomes-${PPID}"
+   ```
+
+2. **Recent sessions and memory**: relevant `claude-mem`/OMX memory observations, session checkpoints, `.omx/notepad.md`, and `.omx/state/**` summaries when present.
+3. **Rollout or release summaries**: changelog entries, release notes, PR summaries, compatibility notes, and post-release follow-up reports that show repeated manual steps.
+4. **Optional Chronicle/history integrations**: Chronicle, shell history, or external task timeline summaries if installed and explicitly available. Do not fail when absent.
+5. **Existing inventory**: compare against `.codex/skills/*/SKILL.md`, `.codex/agents/*`, `templates/.claude/skills/*/SKILL.md`, and generated wiki pages to avoid duplicate or overlapping packaging.
 
-If file doesn't exist or is empty: report "No task outcomes recorded in this session." and stop.
+If no evidence source exists or all sources are empty: report "No recurring workflow evidence found." and stop without creating anything.
+
+Parse JSONL outcome entries when available. Each entry has:
 
-Parse JSONL entries. Each entry has:
 ```json
 {"agent_type": "lang-typescript-expert", "skill": "typescript-best-practices", "description": "Fix type error in auth module", "outcome": "success", "model": "sonnet", "timestamp": "2026-04-05T09:30:00Z", "duration_ms": 15000}
 ```
 
-### Phase 2: Pattern Detection
+For prose evidence, extract only dated or source-attributed observations, for example:
 
-Group outcomes by `(agent_type, skill)` tuple:
+```markdown
+- 2026-05-24, release PR summary: repeated manual wiki parity fixes after adding new guides.
+- 2026-05-22, memory #29064: sync-upstream-release-issues dry-run needed repeated validation after issue-reference parsing changes.
+```
 
+### Phase 2: Recurring Workflow Detection
+
+Group evidence by the workflow being repeated, not only by `(agent_type, skill)` tuple. A workflow can span agents, commands, checklist steps, or release procedures.
+
+```
+Workflow: release-docs-parity-check
+  → evidence_count: 4 dated occurrences
+  → successful_reuse: 3
+  → failure_or_friction: 1
+  → sources: [memory:29148, PR summary, changelog, wiki staleness check]
+  → existing_overlap: wiki, update-docs, sauron-watch
 ```
-Pattern: (lang-typescript-expert, typescript-best-practices)
-  → success: 5, failure: 1, total: 6
-  → success_rate: 0.83
-  → descriptions: ["Fix type error...", "Refactor module...", ...]
+
+Filter qualifying candidates:
+
+- `evidence_count >= threshold` (default: 3), or two strong dated incidents plus high user impact.
+- Repeated manual judgment or sequencing exists; one-off bugs are not enough.
+- Evidence includes dates or source names, not vague recollection.
+- Candidate is not already fully covered by an existing skill, custom subagent, automation, or documented checklist.
+- R006 separation of concerns can be preserved: package one coherent responsibility, not a catch-all meta-agent.
+
+### Phase 3: Build the Shortlist
+
+For each candidate, create a shortlist entry before generating any SKILL.md proposal. Every entry must include these fields:
+
+```markdown
+## Candidate: {candidate-name}
+
+**Workflow**: {one-sentence recurring workflow description}
+**Evidence / Dates**:
+- {date or source}: {specific repeated task, success, failure, or friction point}
+- {date or source}: {specific repeated task, success, failure, or friction point}
+**Frequency / Confidence**: {count and low|medium|high confidence with reason}
+**Recommended Form**: {Skill | Custom subagent | Automation | Skip}
+**Duplicate / Overlap Check**: {existing skills, agents, hooks, scripts, docs, or "none found"}
+**Why**: {why packaging would reduce repeated manual work or improve safety}
+**Why Not**: {risks, overlap, insufficient evidence, or why a lighter form may be better}
 ```
 
-Filter qualifying patterns:
-- `success_count >= threshold` (default: 3)
-- `success_rate >= 0.8`
-- Not already an existing skill (check `.codex/skills/*/SKILL.md`)
+Recommended-form guidance:
+
+| Form | Use When | Do Not Use When |
+|------|----------|-----------------|
+| Skill | A repeatable human-invoked workflow/checklist improves outcomes and needs judgment | Existing skill already covers it or it is fully automatable |
+| Custom subagent | A specialized role with stable responsibilities, tools, and boundaries is recurring | It is just a checklist or would violate R006 by mixing unrelated duties |
+| Automation | The steps are deterministic, cheap to validate, and safe to run without judgment | User approval, external credentials, or destructive actions are required |
+| Skip | Evidence is weak, duplicated, obsolete, or one-off | There is enough dated evidence and clear reuse value |
 
-### Phase 3: Generate Proposals
+### Phase 4: Generate Packaging Proposals
 
-For each qualifying pattern, generate a SKILL.md proposal:
+Only for shortlist entries whose recommended form is `Skill` or `Custom subagent`, generate a proposal. For `Automation`, recommend the script/hook/check location and required guardrails. For `Skip`, explain the evidence gap.
 
 ```markdown
-## Proposal: {proposed-skill-name}
+## Proposal: {proposed-package-name}
 
-**Source Pattern**: {agent_type} + {skill} ({success_count} successes, {success_rate}% rate)
-**Confidence**: {low|medium|high} (based on count and rate)
+**Recommended Form**: {Skill | Custom subagent | Automation | Skip}
+**Source Workflow**: {workflow} ({evidence_count} evidence points, {confidence})
+**Evidence Window**: {earliest date/source} → {latest date/source}
+**Confidence**: {low|medium|high} (based on frequency, recency, and outcome consistency)
 
-### Proposed SKILL.md
+### Proposed Artifact
 
 name: {proposed-name}
-description: {inferred from common description patterns}
+description: {inferred from recurring workflow evidence}
 scope: core
-user-invocable: false
+user-invocable: {true|false}
 
 ### Rationale
-{Why this pattern should be extracted as a skill — based on frequency and success rate}
+{Why this workflow should be packaged — based on dates, frequency, success/failure pattern, and user impact}
 
-### Overlap Check
-{List any existing skills with >50% keyword overlap}
+### Duplicate / Overlap Check
+{List existing skills, agents, hooks, scripts, or wiki docs with meaningful overlap and how this proposal differs}
+
+### Guardrails
+{User approval, R006 responsibility boundary, R020 verification requirement, dry-run behavior, and non-destructive defaults}
 ```
 
 **Confidence scoring**:
-| Successes | Rate | Confidence |
-|-----------|------|------------|
-| 3-5 | >= 0.8 | low |
-| 6-10 | >= 0.85 | medium |
-| 10+ | >= 0.9 | high |
 
-### Phase 4: Present to User
+| Evidence | Recency / Outcome | Confidence |
+|----------|-------------------|------------|
+| 2 strong incidents or 3 weak signals | Mixed outcomes or older than 90 days | low |
+| 3-5 dated occurrences | Mostly successful or repeated friction in last 90 days | medium |
+| 6+ dated occurrences | Clear recurrence, recent evidence, and stable success criteria | high |
 
-Display proposals in ranked order (highest confidence first):
+### Phase 5: Present to User
 
-```
-[skill-extractor] {N} skill candidates detected
+Display the shortlist in ranked order (highest confidence and lowest overlap first):
+
+```text
+[skill-extractor] {N} recurring workflow candidates detected
 
-  1. [high] proposed-skill-name
-     Source: {agent_type} + {skill} (12 successes, 92%)
-     Description: {inferred description}
+  1. [high] release-docs-parity-check
+     Workflow: Validate guide/wiki/template parity before release PRs
+     Evidence: 4 dated sources, latest 2026-05-24
+     Recommended form: Skill
+     Overlap: update-docs, sauron-watch (partial)
+     Why: Prevents repeated CI wiki-staleness failures
+     Why not: May be redundant if update-docs grows the same gate
 
-  2. [medium] another-skill-name
-     Source: {agent_type} + {skill} (7 successes, 86%)
-     Description: {inferred description}
+  2. [medium] upstream-release-issue-validation
+     Workflow: Validate upstream issue references before sync workflows
+     Evidence: 3 dated sources, latest 2026-05-22
+     Recommended form: Automation
+     Overlap: sync-upstream-release-issues script
+     Why: Deterministic validation prevents repeated 404 workflow failures
+     Why not: Better as script/test than human-invoked skill
 
-Select [1-N] to create, "all" to create all, or "skip" to cancel:
+Select [1-N] to create, "all" to create all approved packageable items, or "skip" to cancel:
 ```
 
-### Phase 5: Create Skill (on approval)
+### Phase 6: Create Artifact (on approval only)
 
-Delegate to mgr-creator with the proposal context:
-- Proposed name and description
-- Source pattern data
-- Confidence level
-- Any overlap warnings
+Never create or modify reusable artifacts without explicit user approval after showing the shortlist.
 
-mgr-creator handles: SKILL.md creation, template sync, ontology registration.
+On approval:
+
+- `Skill`: delegate to `mgr-creator` with the full shortlist entry, proposal, overlap warnings, and guardrails.
+- `Custom subagent`: delegate to `mgr-creator` with the R006 responsibility boundary and required related skills/guides.
+- `Automation`: hand off a scoped implementation recommendation; require a dry-run/default-safe mode and R020 verification evidence.
+- `Skip`: record the decision only if the user asks to save it.
+
+mgr-creator handles: SKILL.md creation, template sync, ontology registration, and generated docs parity.
+
+## Recurring-Workflow Packaging Checklist
+
+Before recommending packaging, verify:
+
+- [ ] Evidence is source-attributed and includes dates or stable identifiers.
+- [ ] Frequency meets `--threshold` or has two strong high-impact incidents.
+- [ ] The workflow has stable trigger conditions and a clear stop condition.
+- [ ] Existing skills, agents, hooks, scripts, and wiki docs were checked for duplicate or partial coverage.
+- [ ] Recommended form is justified as `Skill`, `Custom subagent`, `Automation`, or `Skip`.
+- [ ] `Why` and `Why Not` both name concrete evidence.
+- [ ] R006 is preserved: one coherent responsibility and clear boundaries.
+- [ ] R020 is preserved: proposal includes verification evidence or a test/check path.
+- [ ] User approval is required before any artifact creation or mutation.
+- [ ] Dry-run/no-write behavior remains available for review-only usage.
 
 ## Integration
 
 | System | How |
 |--------|-----|
-| task-outcome-recorder | Reads JSONL outcomes as input data |
-| feedback-collector | Complementary: feedback-collector extracts failure patterns, skill-extractor extracts success patterns |
-| mgr-creator | Delegated skill creation on user approval |
+| task-outcome-recorder | Reads JSONL outcomes as one input data source |
+| memory-management / memory-recall | Supplies dated recurring-workflow evidence when available |
+| rollout and release summaries | Surface repeated manual release, docs, and compatibility procedures |
+| optional Chronicle/history | Adds timeline evidence when installed; absence is non-fatal |
+| existing skills and agents inventory | Prevents duplicate skills, subagents, or automations |
+| feedback-collector | Complementary: feedback-collector extracts failure patterns, skill-extractor extracts recurring packageable workflows |
+| mgr-creator | Delegated skill or subagent creation on user approval |
 | skills-sh-search | Check agentskills.io for existing equivalent before creating |
-| R011 (memory) | User Model tracks extraction decisions in Override Decisions |
+| R006 | Enforces coherent responsibility boundaries for custom subagents and skills |
+| R011 (memory) | User Model tracks extraction decisions in Override Decisions when explicitly saved |
+| R020 | Requires verification evidence before completion claims |
 
 ## Hook Integration
 
 The `skill-extractor-analyzer.sh` Stop hook provides a lightweight pre-analysis:
+
 - Reads task outcomes file
-- Counts qualifying patterns
+- Counts qualifying recurring patterns
 - Emits advisory stderr message if candidates found
-- Does NOT create skills (that requires user approval via the skill)
+- Does NOT create skills, subagents, or automation (that requires user approval via the skill)
 
 ## Compatibility Artifact Protocol
 
@@ -152,8 +241,11 @@ Sensitive-path compatibility note: when delegated work touches `.claude/outputs/
 
 ## Safety
 
-- **User approval required**: Never auto-creates skills
-- **Overlap check**: Prevents duplicating existing skills
+- **User approval required**: Never auto-creates skills, subagents, or automation
+- **Evidence-first**: Never recommends packaging from vague memory or unverified anecdotes
+- **Overlap check**: Prevents duplicating existing skills, agents, hooks, scripts, or docs
+- **R006 guardrail**: Rejects catch-all artifacts with mixed responsibilities
+- **R020 guardrail**: Every approved artifact must include a verification path before completion is claimed
 - **Dry-run mode**: Preview without side effects
 - **Advisory hook**: Stop hook is advisory-only (exit 0)
-- **Confidence transparency**: All proposals show confidence scores
+- **Confidence transparency**: All shortlist entries and proposals show confidence scores and evidence dates

package/templates/.claude/skills/structured-dev-cycle/SKILL.md

@@ -93,15 +93,12 @@ A PreToolUse hook in `.codex/hooks/hooks.json` checks this marker and blocks Wri
 └── Output: Implementation complete
 ```
 
-**Codex-Exec Hybrid Option**: When entering Stage 3:
-1. Check `/tmp/.codex-env-status-*` for codex CLI availability
-2. If available AND task involves new file creation → automatically delegate scaffolding to `/codex-exec`:
-   - Display: `[Codex Hybrid] Delegating scaffolding to codex-exec...`
-   - codex-exec generates initial code (strength: fast generation)
-   - Claude expert reviews and refines codex output (strength: reasoning, quality)
-3. If unavailable → display `[Codex] Unavailable — proceeding with Claude experts directly` and proceed with standard implementation via Claude experts
-
-Suitable for codex hybrid: new files, boilerplate, test stubs, scaffolding
+**Optional Codex Plugin Interop**: When entering Stage 3:
+1. Use domain expert agents as the default implementation path.
+2. If the native Claude Code plugin `openai/codex-plugin-cc` is explicitly installed and requested, it may provide Codex interop for new-file scaffolding before expert review.
+3. Otherwise display `[Codex Plugin] Not requested — proceeding with expert agents directly` and proceed with standard implementation.
+
+Suitable for optional plugin interop: new files, boilerplate, test stubs, scaffolding
 Not suitable: modifying existing code, architecture-dependent changes
 
 **Exit criteria**: All planned files created/modified, tests written.
@@ -157,7 +154,7 @@ The stage marker file (`/tmp/.codex-dev-stage`) is read by a PreToolUse hook tha
 For complex tasks, Agent Teams is **preferred** when available (R018):
 - Plan: architect agent
 - Verify: reviewer agent(s) — multi-model-verification via Agent Teams
-- Implement: domain expert agent (+ codex-exec hybrid if available)
+- Implement: domain expert agent (+ optional `openai/codex-plugin-cc` interop only when explicitly installed/requested)
 - Compound: QA agent
 
 When Agent Teams is enabled AND task involves 3+ agents or review→fix cycles, using Agent Teams is MANDATORY per R018.

package/templates/AGENTS.md.en

@@ -110,7 +110,6 @@ NO EXCEPTIONS. NO EXCUSES.
 | `/omcustomcodex:npm-version` | Manage semantic versions |
 | `/omcustomcodex:npm-audit` | Audit dependencies |
 | `/omcustomcodex-release-notes` | Generate release notes from git history |
-| `/codex-exec` | Execute Codex CLI prompts |
 | `/optimize-analyze` | Analyze bundle and performance |
 | `/optimize-bundle` | Optimize bundle size |
 | `/optimize-report` | Generate optimization report |
@@ -134,7 +133,7 @@ project/
 |   +-- hooks/                   # Hook scripts (security, validation, HUD)
 |   +-- contexts/                # Context files (ecomode)
 +-- .agents/
-|   +-- skills/                  # Installed skills (123 directories)
+|   +-- skills/                  # Installed skills (120 directories)
 +-- guides/                      # Reference docs (51 topics)
 ```
 

package/templates/AGENTS.md.ko

@@ -110,7 +110,6 @@ oh-my-customcodex로 구동됩니다.
 | `/omcustomcodex:npm-version` | 시맨틱 버전 관리 |
 | `/omcustomcodex:npm-audit` | 의존성 감사 |
 | `/omcustomcodex-release-notes` | 릴리즈 노트 생성 (git 히스토리 기반) |
-| `/codex-exec` | Codex CLI 프롬프트 실행 |
 | `/optimize-analyze` | 번들 및 성능 분석 |
 | `/optimize-bundle` | 번들 크기 최적화 |
 | `/optimize-report` | 최적화 리포트 생성 |
@@ -134,7 +133,7 @@ project/
 |   +-- hooks/                   # 훅 스크립트 (보안, 검증, HUD)
 |   +-- contexts/                # 컨텍스트 파일 (ecomode)
 +-- .agents/
-|   +-- skills/                  # 설치된 스킬 (123 디렉토리)
+|   +-- skills/                  # 설치된 스킬 (120 디렉토리)
 +-- guides/                      # 레퍼런스 문서 (51 토픽)
 ```
 

package/templates/CLAUDE.md

@@ -102,7 +102,7 @@ oh-my-customcodex로 구동됩니다.
 | 개발 | `/omcustomcodex:goal`, `/dev-review`, `/dev-refactor`, `/sdd-dev`, `/structured-dev-cycle` | 목표 실행, 코드 리뷰, 리팩토링, SDD |
 | 검증 | `/omcustomcodex:sauron-watch`, `/deep-verify`, `/adversarial-review` | R017 검증, 릴리즈 품질, 보안 리뷰 |
 | 릴리즈 | `/pipeline auto-dev`, `/omcustomcodex-release-notes`, `/release-plan` | 자동 개발, 릴리즈 노트 |
-| 리서치 | `/research`, `/scout`, `/deep-plan`, `/omcustomcodex:agora` | 병렬 분석, URL 평가, 연구 계획 |
+| 리서치 | `/research`, `/scout`, `/deep-plan`, `/roundtable-debate` | 병렬 분석, URL 평가, 토론형 검증 |
 | 메모리 | `/memory-save`, `/memory-recall` | 세션 메모리 관리 |
 | 시스템 | `/token-efficiency-audit`, `/pre-generation-arch-check`, `/omcustomcodex:lists`, `/omcustomcodex:status`, `/omcustomcodex:help` | 토큰 효율/아키텍처 감사, 전체 목록, 상태, 도움말 |
 
@@ -119,7 +119,7 @@ project/
 |   +-- hooks/                   # 훅 스크립트 (보안, 검증, HUD)
 |   +-- contexts/                # 컨텍스트 파일 (ecomode)
 +-- .agents/
-|   +-- skills/                  # 스킬 (123 디렉토리)
+|   +-- skills/                  # 스킬 (120 디렉토리)
 +-- guides/                      # 레퍼런스 문서 (51 토픽)
 ```
 

package/templates/CLAUDE.md.en

@@ -112,7 +112,6 @@ NO EXCEPTIONS. NO EXCUSES.
 | `/omcustomcodex:npm-version` | Manage semantic versions |
 | `/omcustomcodex:npm-audit` | Audit dependencies |
 | `/omcustomcodex-release-notes` | Generate release notes from git history |
-| `/codex-exec` | Execute Codex CLI prompts |
 | `/optimize-analyze` | Analyze bundle and performance |
 | `/optimize-bundle` | Optimize bundle size |
 | `/optimize-report` | Generate optimization report |
@@ -133,7 +132,7 @@ project/
 +-- AGENTS.md                    # Entry point
 +-- .codex/
 |   +-- agents/                  # Subagent definitions (50 files)
-|   +-- skills/                  # Skills (123 directories)
+|   +-- skills/                  # Skills (120 directories)
 |   +-- rules/                   # Global rules (22 files)
 |   +-- hooks/                   # Hook scripts (security, validation, HUD)
 |   +-- contexts/                # Context files (4 files)

package/templates/CLAUDE.md.ko

@@ -112,7 +112,6 @@ oh-my-customcodex로 구동됩니다.
 | `/omcustomcodex:npm-version` | 시맨틱 버전 관리 |
 | `/omcustomcodex:npm-audit` | 의존성 감사 |
 | `/omcustomcodex-release-notes` | 릴리즈 노트 생성 (git 히스토리 기반) |
-| `/codex-exec` | Codex CLI 프롬프트 실행 |
 | `/optimize-analyze` | 번들 및 성능 분석 |
 | `/optimize-bundle` | 번들 크기 최적화 |
 | `/optimize-report` | 최적화 리포트 생성 |
@@ -133,7 +132,7 @@ project/
 +-- AGENTS.md                    # 진입점
 +-- .codex/
 |   +-- agents/                  # 서브에이전트 정의 (50 파일)
-|   +-- skills/                  # 스킬 (123 디렉토리)
+|   +-- skills/                  # 스킬 (120 디렉토리)
 |   +-- rules/                   # 전역 규칙 (22 파일)
 |   +-- hooks/                   # 훅 스크립트 (보안, 검증, HUD)
 |   +-- contexts/                # 컨텍스트 파일 (4 파일)

package/templates/README.md

@@ -46,7 +46,7 @@ templates/
 +-- workflows/                        # project-level pipeline definitions
 +-- .claude/
 |   +-- agents/                       # agent definitions (50 files)
-|   +-- skills/                       # skill modules (123 SKILL.md files)
+|   +-- skills/                       # skill modules (120 SKILL.md files)
 |   +-- rules/                        # global rules (22 files)
 |   +-- hooks/                        # hook registry and scripts (40 scripts)
 |   +-- contexts/                     # context files
@@ -65,7 +65,7 @@ The counts below should stay aligned with `templates/manifest.json`, README comp
 
 Flat agent definition files. During Codex installation these land under `.codex/agents/`.
 
-### Skills (123)
+### Skills (120)
 
 `templates/.claude/skills/*/SKILL.md`