codex-plan-review
Skill6 starsChanged 8 days agoMITView on GitHub
Get a second opinion on your current implementation plan by sending it to Codex for independent review. Use after brainstorming, after writing a plan, or when you want to iterate on a plan before execution. Triggers on 'get a second opinion', 'review plan with codex', 'codex review', or 'iterate on plan'.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- ultimatile/development-skillstheir own version, first seen Oct 20, 2026
- dgrissen2/persona-review-kittheir own version, first seen Aug 9, 2026
- tykisgod/quick-questiontheir own version, first seen Mar 28, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- tykisgod/quick-question★ 12same name
- dgrissen2/persona-review-kit★ 0same name
- ultimatile/development-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository22 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- adding-an-add-to-claude-buttonSkill
- bd-to-br-migrationSkill
- codex-reviewSkill
- debugging-kubernetes-rolloutsSkill
- debugging-nodejs-appsSkill
- de-sloppingSkill
- detect-overengineeringSkill
- evaluating-third-party-toolsSkill
- finding-latest-versionsSkill
- git-reconSkill
- humanizerSkill
- hyprland-luaSkill
- mining-claude-sessionsSkill
- minto-pyramidSkill
- nonstopSkill
- operating-frappe-on-kubernetesSkill
- presentation-blueprintSkill
- pr-fixSkill
- pr-landSkill
- prSkill
- watching-pr-bot-reviewsSkill
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 10 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
10
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
If you were dispatched as a subagent to execute a specific delegated task, do NOT follow this skill.
- Distinctive: line 16
After brainstorming produces a design spec
- Distinctive: line 17
After writing an implementation plan
- Distinctive: line 18
When you want a critical review before executing a plan
- Distinctive: line 19
When the user says "get a second opinion" or "review this plan"
- Distinctive: line 31
If no plan file is found, ask the user to point you to the plan or spec they want reviewed.
- Distinctive: line 35
Read the full plan, so you can weigh Codex's findings against it.
- Very specific: line 39
Run the review with codex-review, from the codex-review skill, in the background:
- Distinctive: line 51
After Codex returns its review:
- Distinctive: line 55
Ask: "Want to incorporate any of this feedback into the plan?"
- Distinctive: line 59
If the user wants to update the plan:
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<SUBAGENT-STOP>
If you were dispatched as a subagent to execute a specific delegated task, do NOT follow this skill. Skip it and return to the task you were given.
</SUBAGENT-STOP>Codex Plan Review
Send the current working plan to Codex (gpt-6-astra) for an independent second opinion. This skill is used to iterate on plans before committing to implementation.
When to Use
- After brainstorming produces a design spec
- After writing an implementation plan
- When you want a critical review before executing a plan
- When the user says "get a second opinion" or "review this plan"
Process
1. Find the Current Plan
Locate the most recent plan by checking these locations in order:
- Design specs:
docs/superpowers/specs/*.md(most recent by date prefix) - Plan files:
docs/superpowers/plans/*.md(most recent by date prefix) - Active context:
.claude/memory-bank/activeContext.mdfor references to current work
If no plan file is found, ask the user to point you to the plan or spec they want reviewed.
2. Pick the Context
Read the full plan, so you can weigh Codex's findings against it. Pick the file that holds the project's rules or constraints: CLAUDE.md, or the memory bank file that describes the work.
3. Send to Codex
Run the review with codex-review, from the codex-review skill, in the background:
codex-review plan "$PLAN" --context CLAUDE.mdcodex-review is on PATH where skillsync installed the skills. In a plugin install, call it by its full path and stay in the project directory: <this skill's base directory>/../codex-review/bin/codex-review.
Add --focus "..." for what to check hardest. A high-effort review of a long plan takes several minutes, and you are told when it exits. It prints the review path and a verdict line. Exit 1 means Critical or Major findings. Exit 2 means there is no usable review: read the log tail it prints.
4. Present the Review
After Codex returns its review:
- Present the Codex feedback to the user clearly
- Highlight any blockers or high-severity concerns
- Ask: "Want to incorporate any of this feedback into the plan?"
5. Iterate if Requested
If the user wants to update the plan:
- Edit the plan file with the agreed changes
- Optionally run Codex again for another pass. Ask first, so the review does not loop.
Model
The model pin, its fallback and the dated model table live in the codex-review skill.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.