Skip to content
Agentmart
Submit a tool

adversarial-review-loop

Bounded build-then-independent-review loop for a meaningful code change.

Skill by petralabx

IndexedLicense unknown
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/petralabx-adversarial-review-loop

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Bounded build-then-independent-review loop for a meaningful code change. Plans, implements, verifies, and ships a pull request, then runs an INDEPENDENT adversarial reviewer (a different model family than the builder, Codex-class for general review) that returns severity-ranked findings; the builder fixes every finding at or above a configured severity threshold, pushes, and re-reviews. Persists plan, branch, PR, findings, verdict, and iteration so the run resumes safely. Stops only when the reviewer approves, only sub-threshold findings remain, or max-iter is reached — and NEVER reports a stalled, errored, or exhausted run as approved. Use for code changes that benefit from an independent reviewer and several structured review-and-fix rounds — e.g. "clodex", "adversarial review loop", "review and fix until approved", "independent code review loop".

Use cases: agent-building

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/adversarial-review-loop

What's inside

Skill name
adversarial-review-loop
Description
Bounded build-then-independent-review loop for a meaningful code change. Plans, implements, verifies, and ships a pull request, then runs an INDEPENDENT adversarial reviewer (a different model family than the builder, Codex-class for general review) that returns severity-ranked findings; the builder fixes every finding at or above a configured severity threshold, pushes, and re-reviews. Persists plan, branch, PR, findings, verdict, and iteration so the run resumes safely. Stops only when the reviewer approves, only sub-threshold findings remain, or max-iter is reached — and NEVER reports a stalled, errored, or exhausted run as approved. Use for code changes that benefit from an independent reviewer and several structured review-and-fix rounds — e.g. "clodex", "adversarial review loop", "review and fix until approved", "independent code review loop".
License
unknown
Instructions
156 lines in SKILL.md · view on GitHub

Files (2)

  • SKILL.md
  • reference.md

The license isn't open source, so the instructions aren't shown here. Read SKILL.md on GitHub

More from petralabx/skills

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Jul 26, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
License unknown
Repository
github.com/petralabx/skills
First seen
Oct 24, 2026
Last checked
Oct 24, 2026
Source
GitHubgithub-graphql

More from petralabx

Cite this page

adversarial-review-loop skill. Agent Mart. https://agentmart.dev/skills/petralabx-adversarial-review-loop (last updated 2026-10-24)