self-critic

self-critic is a skill for Claude Code from Rockielab/rockie-claude. It costs 40 tokens per session (304 once invoked), scanned A, original, Apache-2.0.

An internal review loop that repeatedly checks a draft with a fresh adversarial critic. It is configured to turn merged pull requests into release notes grouped as Added, Changed, and Fixed.

In plain words
What is it for?
Use it to review release-note drafts, apply fixes identified by the critic, and verify that each pull-request change has one clear, user-facing line.
Why use it?
It looks for serious problems in the draft before it is returned and requires two consecutive clean reviews. This reduces the chance of shipping unclear or incorrect release notes.

Skill for Claude Code

Written for Claude Code: installed under .claude/. Also seen: mentions subagents.

Part of the rockie-claude plugin — 29 skills, 1 MCP server shipped together

Good fit Use it to review release-note drafts, apply fixes identified by the critic, and verify that each pull-request change has one clear, user-facing line.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/rockielab/rockie-claude/self-critic
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add Rockielab/rockie-claude --skill self-critic
Clone the repo
git clone --depth 1 https://github.com/Rockielab/rockie-claude

Made for: Claude Code.

Or install rockie-claude, the plugin that ships this one along with the rest of its 29 skills, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for self-critic

README.md
[![agentmods](https://agentmods.dev/badge/skills/rockielab/rockie-claude/self-critic/github.svg)](https://agentmods.dev/skills/rockielab/rockie-claude/self-critic)
Your own site
<a href="https://agentmods.dev/skills/rockielab/rockie-claude/self-critic"><img src="https://agentmods.dev/badge/skills/rockielab/rockie-claude/self-critic/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for self-critic

Your own site · 80×15
<a href="https://agentmods.dev/skills/rockielab/rockie-claude/self-critic"><img src="https://agentmods.dev/badge/skills/rockielab/rockie-claude/self-critic.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 40 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 304 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00040 $0.00304
Opus 5 $0.00020 $0.00152
Sonnet 5 $0.00008 $0.00061
Haiku 4.5 $0.00004 $0.00030

Measured 11d ago against content hash a8f0057feac8, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

self-critic scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (critic_loop.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

project-harness/skills/agent-builder/examples/release-notes-writer/.claude/skills/self-critic/SKILL.md · 31 lines

What it actually says

self-critic — the agent's internal adversarial flow

This skill is wired in by default (the agent-builder is biased to emit self-adversarial agents). Before this agent returns ANY deliverable for the goal below, harden it here.

Goal under guard

Turn a list of merged pull requests into a clear, user-facing release-notes section grouped by Added / Changed / Fixed, with one plain-language line per change.

How to run

  1. Produce a draft deliverable.
  2. Run critic_loop.py with the domain critic (domain-critic.md) as a fresh, no-memory reviewer of the draft.
  3. On any CRITICAL: apply the cited fixes, re-run. The clean-pass counter resets to zero on any failed round.
  4. Converge only on TWO consecutive zero-CRITICAL passes (cap: MAX_ROUNDS=6). Never return a deliverable that has not converged.

Prefer the subagent mode (a fresh subagent per round). When the agent cannot spawn subagents, use the single-process fallback with a hard context reset between rounds — weaker isolation, same termination rules.

Prove the loop logic any time with critic_loop.py --selftest.

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 31 lines · 40 tokens per session scan A a8f0057feac8

Subscribe to this mod's changes

self-critic is a skill published in the GitHub repository Rockielab/rockie-claude (21 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 40 tokens to every session and 304 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

pi-parity-review

Adversarially verify that a ported change is faithful to the original pi implementation (TS source + published npm build). Use after porting upstream pi changes, or standalone on any area of this repo ("is X faithful to pi?").

sky-valley/pi · 54 tokens

pi-go-review

Review ported Go code for idiomatic quality — that the port maximizes Go rather than transliterating TypeScript. Use after porting upstream pi changes, or standalone on any diff in this repo.

sky-valley/pi · 45 tokens

repo-asset-stocktake

Audit a project repo's non-code assets — tool configs, CI/GitHub workflows, runbooks, other docs — for assets whose consumer has vanished, and assign Keep/Update/Retire/Merge verdicts. Use when the user says "audit my repo assets", "which configs/workflows/runbooks are dead", "repo asset stocktake"…

shimo4228/claude-harness · 156 tokens

review-to-lint

A procedure for turning repeatable review checks into scripts, while leaving meaning and judgment to the reviewer. It can use a reviewer's checklist or past review records to find checks that a program can decide.

shimo4228/claude-harness · 213 tokens

codex-review

Cross-model second opinion from the OpenAI Codex CLI (a different model family), read-only, in two seams — (1) code review of the current diff; (2) plan-stage premise challenge of a design packet (refute / missing / alternative, never a design). Opt-in only — use ONLY when the user explicitly asks with "codex review"…

shimo4228/claude-harness · 178 tokens

auto-run

Autonomous personalized research loop. Use when the user wants to research a topic autonomously, run a research loop, start adaptive research, or use presets like technique-scout or cross-domain. Triggers on: 'auto run', 'research loop', 'autonomous research', 'run research', 'start research', 'adaptive research'.

primeline-ai/claude-adaptive-research · 70 tokens