How do I install the Critique skill?

Run npx skillmds@latest add raven-sourav/critique in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Critique skill do?

Skill: Critique a Draft It is listed under Coding & Dev Tools on SkillMD.

Is Critique safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Critique?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Critique free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Critique?

raven-sourav (@raven-sourav) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Critique by raven-sourav — Agent Skill

Name: Critique
Author: raven-sourav

File contents

Skill: Critique a Draft

Trigger

User submits a draft for critique, or critique runs as part of generation.

Process

Step 1: Load references

Always load: references/rubber-duck-escalator.md
Load for voice check: data/brain/pmm_brain.json (voice_profile_sourav, quality_bar_examples)

Step 2: Run 5 phases sequentially

Each phase sees the draft + all prior phase results:

MIRROR — Is this lived experience or knowledge-reporting?
PROBE — What uncomfortable truth is being dodged?
CHALLENGE — Where is the logic soft? Evidence check.
ILLUMINATE — What threads haven't been connected?
CRYSTALLIZE — Does it sound like Sourav specifically?

Step 3: Score and gate

Score each phase 1-10 (critique first, score based on own critique)
All 5 must be 8+ to pass
If any < 8: provide specific rewrite instructions
If composite < 4.0: flag for fundamental rethinking

Step 4: Present scorecard

Show phase-by-phase results with specific line references.

Output Format

━━━━ RUBBER DUCK SCORECARD ━━━━

MIRROR:      [score]/10 — [one-line summary]
PROBE:       [score]/10 — [one-line summary]
CHALLENGE:   [score]/10 — [one-line summary]
ILLUMINATE:  [score]/10 — [one-line summary]
CRYSTALLIZE: [score]/10 — [one-line summary]

COMPOSITE: [score] | STATUS: [PASS/REWRITE/FLAG]

[If REWRITE: specific instructions ranked by severity]
[If FLAG: what needs fundamental rethinking]

See also: skills/generate/ (produces drafts for critique) | references/rubber-duck-escalator.md (full protocol)

See also: skills/generate/ (revision loop with critique) | references/rubber-duck-escalator.md (full protocol) | skills/_index.md for full data flow

Frequently asked questions

Critique