Skip to content
Back to skills

Superreview

ASecurity

Review a change set or explicit project scope for intent, correctness, safety, reuse, and maintainability; independently validate findings and produce one report without applying fixes.

  • 32 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 1, 2026
ai-agents

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned October 1, 2026

npx -y skills add kochetkov-ma/claude-brewcode --skill superreview --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Superreview?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Superreview
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/kochetkov-ma-superreview/badge)](https://www.skillsdirectory.com/skills/kochetkov-ma-superreview)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: superreview
description: Review a change set or explicit project scope for intent, correctness, safety, reuse, and maintainability; independently validate findings and produce one report without applying fixes.
---

# Superreview

Use the native Codex review workflow in [references/review-contract.md](references/review-contract.md). This skill authorizes review and report artifacts only: never edit reviewed code, instructions, agents, skills, or configuration, including this skill itself. Suggest corrections separately.

Resolve the requested scope before work. Explicit paths, exclusions, revisions, and user narrowing win. Default to staged + unstaged + untracked changes; with a clean tree review the last two commits, falling back to one or the root commit. QUICK is the default; requests for depth, completeness, pre-merge, or release review select EXTENDED. Depth changes effort, never scope or mutation authority. Announce the concrete scope, depth, baseline, checks, and expected report in English.

Read applicable `AGENTS.md` instructions and their referenced rules explicitly; `.codex/rules/` and `paths:` metadata do not load rules automatically. Reuse project check commands and live native reviewer roles. Parse agent TOML as TOML, not YAML; availability comes from the current runtime. Do not create agents or require an `intent-guard` role.

Review in order: simplification/reuse assessment, correctness/safety, then maintainability/clarity. Simplification is a report pass, not permission to run a mutating simplify workflow. At both depths compare requested and delivered behavior using the intent procedure in the contract. EXTENDED adds bounded domain reviews and independent validation of every candidate; QUICK retains concrete evidence and independent validation for material findings and states its narrower coverage.

Write one deduplicated report at `.codex/reports/<YYYYMMDD-HHMMSS_superreview>/REPORT.md`, including findings, validation, checks, intent coverage, and limitations. Summarize the actionable result in the conversation language. Report missing checks/roles and incomplete validation explicitly; zero findings with missing coverage is not a clean review.

Files in this skill

  • SKILL.md2.2 KB
  • references/review-contract.md8.6 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.