Forces an adversarial mindset on a plan, PRD, design, or code change — break it rather than bless it. Loaded at the /dr-plan checkpoint and /dr-qa.
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 — Break It Before Production Does
The default failure mode of review is confirmation: a reader who wants the
artifact to be correct finds reasons it is. This skill inverts the stance. Your
job is to construct the strongest case that the artifact is **wrong,
incomplete, or unsafe** — and only then weigh that case against its defence.
This is the *general* adversarial lens. Two sibling skills sharpen specific
axes: `edge-case-hunter` (boundary and failure inputs the artifact does not
handle) and `structure-review` (organisation, completeness, internal
contradiction). Load whichever the task calls for; this skill covers the
mindset and the whole-artifact attack.
## The core rule: "no findings" is an alarm, not an all-clear
If an adversarial pass produces zero findings, that is evidence the pass was
shallow — not that the artifact is flawless. Before you may record "no
material findings", you MUST have done at least one of:
- named the artifact's single most load-bearing assumption and tried to falsify it;
- traced one concrete failure scenario end-to-end and shown it cannot occur;
…
The line at the top of this page pins the label git:20261004.8dfccd1, 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_ycpyrgisxhj7kwxs
GET https://markdownregistry.com/api/v1/resolve?ref=arcanada-one/datarim/adversarial-review
GET https://markdownregistry.com/api/v1/blob/1526b3ff7b44b61f3df76b57df8a145072611bc578700b207b41d36e95f202de
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.
arcanada-one/datarim · plugins/product-discovery/skills/decision-record/SKILL.md · Record bounded product decisions with counterevidence, uncertainty, and explicit reversal conditions.
arcanada-one/datarim · plugins/product-discovery/skills/hypothesis-validation/SKILL.md · Use when a product decision depends on an uncertain claim and needs a bounded, falsifiable check.
arcanada-one/datarim · skills/brainstorming/SKILL.md · MUST use this before any creative work — features, components, functionality, behavior changes. Explores user intent…
arcanada-one/datarim · skills/consilium/SKILL.md · Assemble a panel of Datarim agents for multi-perspective analysis of decisions. Use for /dr-design stage (L3-4) or…
arcanada-one/datarim · skills/context-window-self-clearing/SKILL.md · Preserve Datarim task continuity while orchestrated Claude Code or Codex sessions compact or clear context at…
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…
syahiidkamil/software-engineer-ai-agent-atlas · .claude/skills/adversarial-review/SKILL.md · Adversarially review something just built — presume it is broken and find where. Use after implementing a feature…
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…
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…
