Codex Plan Review
Purpose
Use this skill to adversarially review a plan before implementation starts.
When to Use
Use after creating a plan but before implementing code.
Typical workflow:
- Create plan (plan.md) → 2. /codex-plan-review → 3. Refine plan → 4. Implement → 5. /codex-impl-review
Not a substitute for code review. Reviews plan quality, not implementation.
Prerequisites
-
A Markdown plan file exists (e.g. plan.md ) with headings for sections, steps, or phases.
-
codex CLI is installed and authenticated.
-
codex-review skill pack is installed (npx github:lploc94/codex_skill ).
Runner
RUNNER="{{RUNNER_PATH}}"
Workflow
-
Ask user to choose debate effort level: low , medium , high , or xhigh (default: high ). Set EFFORT .
-
Run pre-flight checks (see references/workflow.md §1.5).
-
Build prompt from references/prompts.md (Plan Review Prompt ), following the Placeholder Injection Guide.
-
Start round 1 with node "$RUNNER" start --working-dir "$PWD" --effort "$EFFORT" .
-
Poll with adaptive intervals (Round 1: 60s/60s/30s/15s..., Round 2+: 30s/15s...). After each poll, report specific activities from poll output (e.g. which files Codex is reading, what topic it is analyzing). See references/workflow.md for parsing guide. NEVER report generic "Codex is running" — always extract concrete details.
-
Parse Codex issues (ISSUE-{N}
- VERDICT ) using references/output-format.md .
-
Apply valid fixes to the plan, save the plan file, rebut invalid points, and resume with --thread-id .
-
Repeat until APPROVE , stalemate, or hard cap (5 rounds).
-
Return final debate summary, residual risks, and final plan path.
Effort Level Guide
Level Depth Best for Typical time
low
Surface check Quick sanity check ~2-3 min
medium
Standard review Most day-to-day work ~5-8 min
high
Deep analysis Important features ~10-15 min
xhigh
Exhaustive Critical/security-sensitive ~20-30 min
Required References
-
Detailed execution steps: references/workflow.md
-
Prompt templates: references/prompts.md
-
Output contract: references/output-format.md
Rules
-
If Claude Code plan mode is active, stay in plan mode during the debate. Otherwise, operate normally.
-
Do not implement code in this skill.
-
Do not claim consensus without explicit VERDICT: APPROVE or user-accepted stalemate.
-
Preserve traceability: each accepted issue maps to a concrete plan edit.