agentleFS
ChianWChianW/C31/skills/review/C31-review-security/SKILL.md

C31-review-security

Skill1 starsChanged 3 months agoMITView on GitHub

Output language: Respond automatically in the user's conversation language. Return exactly one JSON object: Rules: - autofix_class: safe_auto only for adding input sanitization or escaping. Auth/logic fixes are gated_auto or manual. - confidence: 100 = I can write a curl to exploit this; 75 = clear vulnerable pattern; 50 = plausible with specific input; 25 = suspicious but unlikely. - No findings = empty arrays. Do not cry wolf.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read-only — does not edit files.

    line 7
  2. Distinctive:

    Read the intent summary and diff.

    line 44
  3. Distinctive:

    Check if input reaches a query builder, shell, template, or eval.

    line 47
  4. Distinctive:

    Check auth guards: is ownership verified?

    line 48
  5. Distinctive:

    Check output construction: is user input reflected in HTML/JSON without escaping?

    line 49
  6. Distinctive:

    Check for secrets in strings, config, or logs.

    line 50
  7. Distinctive:

    No findings = empty arrays.

    line 82

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Multilingual Triggers

LanguageTrigger phrases
ENC31-review-security
ZH安全审查, 安全检查
JAセキュリティレビュー, 脆弱性チェック

Output language: Respond automatically in the user's conversation language.

C31 Review — Security

Focus

  • Injection (SQL, NoSQL, command, template, LDAP)
  • XSS / CSP bypasses / unsafe HTML construction
  • Auth bypass / privilege escalation / missing ownership checks
  • Secrets in code or logs (tokens, passwords, keys)
  • Insecure deserialization / unsafe eval
  • Missing rate limiting / brute-force exposure
  • CORS misconfiguration / unsafe headers
  • Data exposure (PII leak, over-fetching, missing field-level auth)

Severity Calibration

LevelMeaningExamples
P0Direct exploit, data breach, auth bypassSQL injection in public endpoint
P1Elevated privilege or significant data leakMissing ownership check on delete
P2Defense-in-depth gap, narrow exploit windowNo rate limit on login
P3Logging of sensitive data, minor header issuePassword logged in error message

Workflow

  1. Read the intent summary and diff.
  2. For each endpoint, handler, or input boundary touched:
    • Identify all user-supplied input (query params, body, headers, path segments).
    • Check if input reaches a query builder, shell, template, or eval.
    • Check auth guards: is ownership verified? Is the check on the right layer?
    • Check output construction: is user input reflected in HTML/JSON without escaping?
    • Check for secrets in strings, config, or logs.
  3. Flag any exploitable path or defense gap.

JSON Output Contract

Return exactly one JSON object:

{
  "reviewer": "security",
  "findings": [
    {
      "title": "One-line description",
      "severity": "P0|P1|P2|P3",
      "file": "relative/path",
      "line": 42,
      "confidence": 0|25|50|75|100,
      "autofix_class": "safe_auto|gated_auto|manual|advisory",
      "owner": "review-fixer|downstream-resolver|human",
      "requires_verification": true|false,
      "pre_existing": true|false,
      "suggested_fix": "If known, brief fix description"
    }
  ],
  "residual_risks": ["Architecture-level risk 1"],
  "testing_gaps": ["Missing security test for auth bypass"]
}

Rules:

  • autofix_class: safe_auto only for adding input sanitization or escaping. Auth/logic fixes are gated_auto or manual.
  • confidence: 100 = I can write a curl to exploit this; 75 = clear vulnerable pattern; 50 = plausible with specific input; 25 = suspicious but unlikely.
  • No findings = empty arrays. Do not cry wolf.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.