Adversarially review something just built — presume it is broken and find where. Use after implementing a feature, finishing a build, or before shipping, or whenever ATLAS or Boss wants a hostile second opinion on a diff, a running app, or a whole codebase. Spawns clean-context reviewers with opposing lenses (correctness, security, empty-world) and reproduces every finding by running the target before believing it. Distinct from /code-review, which statically reads a diff and never runs the app.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Adversarial review — presume it broken, find where
A normal review reads code and looks for reasons to approve it. An adversarial review starts from the
opposite prior: **this was just built, it is presumed broken, and the job is to find where.** The
builder believes in their work — that belief is exactly the blind spot. So the reviewer must be
incentivized to find failure, not confirm success, and must work *without the builder's context* so it
cannot inherit the builder's assumptions.
This skill is the canonical adversarial-review protocol for ATLAS. Reach for it autonomously after
building anything non-trivial, before shipping, or when handed unfamiliar code to harden.
## When to use it — and which review
- **Use this** when you want the thing *attacked*: a feature you just finished, a build before
delivery, a flow you suspect is fragile, an unfamiliar codebase you must trust.
- **Use `/code-review` instead** for a fast, conservative pass over a **diff or PR** — it statically
reads the change for bugs and cleanups, scores confidence, and **never runs the app**. That is the
right tool for routine change review.
…
The line at the top of this page pins the label git:20260616.dc49c8f, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
GET https://markdownregistry.com/api/v1/artifacts/art_mi7y5bdhqrsks6io
GET https://markdownregistry.com/api/v1/resolve?ref=syahiidkamil/software-engineer-ai-agent-atlas/adversarial-review
GET https://markdownregistry.com/api/v1/blob/a906fc3404c7159c224ba2b2be9c60fbb49652393fdc5a2535de59079185bf3c
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/deploy/SKILL.md · Take a project from working locally to running in production, provider-agnostic — write the DEPLOY.md runbook, gate a…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/find-skills/SKILL.md · Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/free-will/SKILL.md · Deliberate-choice procedure for a medium-to-high-stakes engineering fork — when the first plausible solution (the…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/frontend-design/SKILL.md · Guidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/human-writing/SKILL.md · Write content in any language indistinguishable from a skilled human writer. Avoids AI-sounding phrases, varies…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/learning-from-mistakes/SKILL.md · Record a hard-won engineering lesson after a bug or mistake is resolved. Use when a bug resisted ATLAS until Boss's…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/playwright-cli/SKILL.md · Automate browser interactions, test web pages and work with Playwright tests.
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/skill-creator/SKILL.md · Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/supabase-postgres-best-practices/SKILL.md · Postgres best practices maintained by Supabase, for Postgres running anywhere. Load this skill BEFORE writing or…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/super-product-owner/SKILL.md · Decide what is worth building and what "great" means for it — interview the user one question at a time, write a…
basicmachines-co/basic-memory · .agents/skills/adversarial-review/SKILL.md · Cross-vendor adversarial code review of the current branch. Two different model families (Claude + Codex/GPT) review…
timharris707/skills · skills/run/adversarial-review/SKILL.md · Run an adversarial review of a change before it ships: isolated finders trying to break it, a skeptic pass that kills…
timharris707/skills · plugins/clickai-codex/skills/run/adversarial-review/SKILL.md · Before committing substantial work, at worker completion, before external review, or on an explicit…
arcanada-one/datarim · skills/adversarial-review/SKILL.md · Forces an adversarial mindset on a plan, PRD, design, or code change — break it rather than bless it. Loaded at the…
robertoecf/adversarial-review · skills/adversarial-review/SKILL.md · Use for requested adversarial review of plans, diffs, configs, prompts or skills, and whenever a project requires that…
