Critic
One-shot adversarial plan review. Reads the active plan file or last assistant message and returns 3-10 severity-rated findings covering scope creep, symptom-vs-root-cause, missing tests, fabricated facts, and unverified assumptions. Wraps oh-my-claudecode:critic agent. Triggers "/critic", "review my plan", "plan critique", "adversarial review".
Install
$ agentstack add skill-ao92265-claude-code-playbook-critic ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-ao92265-claude-code-playbook-critic) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
/critic — Plan Review Wrapper
Adversarial second-opinion pass on a plan or proposal before execution begins.
When to invoke
- Plan file written, before ExitPlanMode approval
- Long autopilot/ralph prompt drafted, before launch
- User asks "is this plan any good"
- After a deep-interview produced a spec, before implementation
When NOT to invoke
- Trivial single-file changes
- Bug fixes where the fix is mechanical
- During implementation (use
code-reviewerinstead)
Workflow
- Identify target:
- If active plan file exists at
~/.claude/plans/*.md(latest mtime) → use it - Else use the last assistant message in conversation
- User can pass explicit path:
/critic
- Spawn
oh-my-claudecode:criticagent (model=opus) with prompt:
``` Review the attached plan. Find 3-10 specific issues across these axes:
- Scope creep (does it sneak in unrelated changes?)
- Symptom vs root cause (is it patching a symptom?)
- Missing tests / verification gaps
- Fabricated facts (file paths, function names, APIs that may not exist)
- Unverified assumptions about codebase state, user intent, library behavior
- Sequencing risks (steps that depend on prior unproven steps)
Format each finding:
- [SEVERITY: HIGH|MED|LOW]
- Location:
- Why it matters:
- Recommended fix:
Max 10 findings. Return only the findings. No preamble, no summary. ```
- Surface findings verbatim to user. Do NOT auto-edit the plan.
Hard Rules
- Critic agent is READ-ONLY. No file edits.
- Findings include file:line where applicable.
- Severity is honest, no padding to hit a count.
- If plan is genuinely solid: return 0–2 findings + explicit "no major concerns" line. Don't manufacture issues.
Triggers
/critic, "review my plan", "plan critique", "adversarial review", "second opinion on plan"
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: ao92265
- Source: ao92265/claude-code-playbook
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.