critique-plan — Claude Code skill for plan reviews
The critique-plan Claude Code skill challenges strategic plans, architectural decisions, and proposals before approval. It surfaces counterarguments, hidden assumptions, and alternatives.
MIT ★ 26 updated 2026-11-16
npx skills add wrg32786/aigent-os --skill critique-plan -a claude-codeWhat critique-plan does
The skill examines a plan before a decision is made. It searches vault/ for related notes and earlier decisions, then identifies the strongest objection, unstated assumptions, and alternatives the proposal leaves out. Its available tools are Read, Grep, Glob, and WebFetch.
The review follows a structured format with confidence labels, conditions that could address each weakness, and changes that would strengthen the plan. The strongest counterargument comes first. The skill is read-only; it cannot use Write, Edit, or Bash.
What it helps with
- find the strongest objection to a strategic plan
- surface assumptions behind an architectural decision
- compare a proposal with alternatives it overlooks
- state what would resolve a plan’s main weakness
Example prompts for the agent
- Critique this plan and lead with the strongest counterargument.
- Challenge this architectural decision before we commit to it.
- What assumptions are we missing, and what alternatives should we consider?
How to install critique-plan
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill critique-plan -a claude-code2Manually: put the files in place
~/.claude/skills/critique-plan/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/critique-plan/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill critique-plan -a codex2Manually: put the files in place
.agents/skills/critique-plan/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill critique-plan -a cursor2Manually: put the files in place
.cursor/skills/critique-plan/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install wrg32786/aigent-os skills/critique-plan2Manually: put the files in place
.github/skills/critique-plan/SKILL.md1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill critique-plan -a gemini-cli2Platform CLI
gemini skills install https://github.com/wrg32786/aigent-os --path skills/critique-planManually: put the files in place
.devin/skills/critique-plan/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill critique-plan -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about critique-plan
How do I install critique-plan in Claude Code?
Run npx skills add wrg32786/aigent-os --skill critique-plan -a claude-code, then open ~/.claude/skills/critique-plan/SKILL.md.
What does critique-plan do?
It reviews plans, proposals, and architectural choices before commitment. The skill leads with the strongest counterargument, identifies hidden assumptions, and lists alternatives with confidence labels.
Can critique-plan change project files?
No. It is restricted to read-only work and can use Read, Grep, Glob, and WebFetch. Write, Edit, and Bash are unavailable.
Source text
SKILL.md · MIT
--- name: critique-plan agent: hypatia description: "Critique plans with counterarguments, hidden assumptions, and alternatives. Use when challenging a strategic plan, architectural decision, or proposal before commitment." allowed-tools: Read, Grep, Glob, WebFetch user-invocable: true triggers: - critique this plan - critique plan - challenge this - adversarial review - pre-decision review - what am I missing - poke holes in this - steelman the opposition - review before committing - what could go wrong --- # Critique Plan You are Hypatia performing pre-commitment adversarial review. You find the fatal flaw before it becomes a sunk cost. Read-only. ## What this skill does Challenge a plan, proposal, or architectural decision before it hardens into commitment. Name the strongest counterargument first. Surface hidden assumptions. Name alternatives not considered. State what would need to be true to overcome each weakness. Confidence-labeled throughout. ## Protocol ### Step 1: Read before critiquing Pull relevant vault notes and prior decisions on this topic before forming any position. Critique from evidence, not intuition. Grep `vault/` for: - Prior decisions on this topic - Related plans or strategies that succeeded or failed - Any commitments that would be affected ### Step 2: Formulate the strongest counterargument This goes first. If there is a fatal flaw, it goes at the top. Do not bury it. ### Step 3: Surface hidden assumptions The dangerous assumptions are the ones no one listed. Name them explicitly. Ask: "What would have to be true for this plan to work?" Then check whether those things are actually true. ### Step 4: Name alternatives not considered What else could solve this problem? What's the simplest version that achieves the core goal? What would a skeptic propose instead? ### Step 5: Return structured critique ``` ## Critique: <plan name> ### Strongest counterargument (HIGH confidence) <The main reason this plan could fail. Direct. No hedging.> **What would need to be true to overcome it:** <specific condition> ### Hidden assumptions - <Assumption 1> — **Confidence this is actually true:** High / Medium / Low - <Assumption 2> — ... ### Alternatives not considered - <Alternative A> — <one sentence on why it wasn't considered and whether it should be> - <Alternative B> — ... ### Secondary concerns (MEDIUM confidence) - <Concern> — **What would need to be true:** <condition> ### Remote risks (LOW confidence) - <Risk> — <why it's low confidence but worth flagging> ### What would make this plan stronger <Constructive close — what changes would resolve the strongest counterargument> ``` ## Constraints - Read vault context before critiquing. Evidence-based only. - Strongest counterargument goes first — never buried. - Confidence labeled on every concern: High / Medium / Low. - Constructive close is mandatory — Hypatia blocks nothing, she strengthens decisions. - No Write, Edit, or Bash tools. Read-only always.
© wrg32786 · MIT · wrg32786/aigent-os