Skip to content
Skills Library
Security · Review & audit

attack

/attack

Adversarial review of a change to automation code. Three parallel reviewers try to break it through trust and injection, failure and rerun, and contract changes, and a finding survives only if a run demonstrates it. Use for "attack this" or "adversarial review", and before merging anything that touches credentials, deploy targets, trigger rules, or shared templates.

Repository
ian ↗
Licence
MIT
Last repository change
09.11.2026

Assessment and security

86/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content21 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 09.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Adversarial review of a change to automation code. Three parallel reviewers try to break it through trust and injection, failure and rerun, and contract changes, and a finding survives only if a run demonstrates it. Use for "attack this" or "adversarial review", and before merging anything that touches credentials, deploy targets, trigger rules, or shared templates.

Triggers from the description

attack thisadversarial review
Original author description

Adversarial review of a change to automation code. Three parallel reviewers try to break it through trust and injection, failure and rerun, and contract changes, and a finding survives only if a run demonstrates it. Use for "attack this" or "adversarial review", and before merging anything that touches credentials, deploy targets, trigger rules, or shared templates.

How it works

SKILL.md has no level-two headings · 16 lines, 309 words

SKILL.md preview
---
name: attack
description: Adversarial review of a change to automation code. Three parallel reviewers try to break it through trust and injection, failure and rerun, and contract changes, and a finding survives only if a run demonstrates it. Use for "attack this" or "adversarial review", and before merging anything that touches credentials, deploy targets, trigger rules, or shared templates.
---

# Attack

Reviewers try to break the change. Every finding needs a run that demonstrates it. Report findings; change nothing unless asked.

1. **Scope.** Use what the user named. Otherwise use the diff from the merge base to the working tree. State the change's intent in two sentences, taken from the commits or the merge request text.
2. **Spawn three reviewers in one message.** Use the `ian:worker` agent, each with `isolation: "worktree"`. Give each one the scope, the intent, one lens, the brief below, and the router's path: `${CLAUDE_PLUGIN_ROOT}/skills/work/SKILL.md`.
   - **Trust.** Who controls each input: environment, CI variable, trigger payload, file, API response? Can a caller change where something deploys, what runs, or whether a check runs? Does any value reach a shell, a path, or a URL unvalidated? Follow each trusted value through everything it resolves into.
   - **Failure.** What happens on a crash at each step, on a rerun, on empty input, on a timeout, or on a partial API failure? Is there any path that reports success without doing the work?
   - **Contract.** What does each consumer see differently: inputs, defaults, exit codes, outputs, artifact names, resource addresses?

   The brief: "Find where this change fails. For every finding, run something real that shows it: a failing test, a script, or a plan. Give each finding a title, the command you ran, what it printed, and what the intent says it should print. Omit anything you did not run, including style comments. Finding nothing is an acceptable outcome."
3. **Verify.** Rerun every reported command yourself. Keep a finding only if its output demonstrates a failure that contradicts the intent. Combine findings with the same root cause.

**Report:** the intent. Then each finding with its command, output, why it contradicts the intent, and which lens found it. Then the number of findings dropped per lens, with the reasons.

Versions

The database has no other skills named attack.

Database snapshot: 13.11.2026 · package version 48acf8358d53 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 95adversarial-review
    Try to break a change or design and keep only findings proven by a run. Fresh adversary agents in their own worktrees attack the change against a restated intent; every finding is rerun before it survives. Also covers blast radius (what a change breaks elsewhere, with a proof…
  • 90attack-shape
    Use when the user wants adversarial stress-testing of a proposed architecture, structure, or shape. Not for tasks that require source or remote-system changes.
  • 100security-review
    Review a change for secrets, injection, supply-chain, and unsafe patterns before merge. Use after any implementation, before merging or releasing, and whenever code touches auth, secrets, data, or external input. Only invoke via the orcinus router.
  • 96security-review
    Use when the user asks for a security review or audit of code or changes (injection, secrets, unsafe input handling, permissions).
  • 90interrogate
    Use for an explicit request to interrogate code, run an adversarial review, or use multiple models to find failures. Reviewers run reproductions, and the lead keeps only failures that a rerun confirms.
  • 100security-review
    Security-focused code review with attack surface mapping and risk classification. Use when reviewing PRs for security, auditing code changes, or analyzing potential vulnerabilities. Triggers on: 'security review', 'use security mode', 'audit this', 'check for vulnerabilities',…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.