agentleFS
sxntixgosxntixgo/frugal-skills/skills/security-review/SKILL.md

security-review

Skill0 starsChanged 4 months agoMITView on GitHub

Reviews code for security vulnerabilities — injection, broken auth/authz, secrets exposure, crypto misuse, SSRF, deserialization, and more — with an exploit scenario and fix per finding. Use when a change touches auth, input handling, data access, file/network I/O, secrets, or before shipping security-sensitive code.

192 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use linkerd-patterns, dependency-cve-sweep and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • xfstudioSkill

    linkerd-patterns

    Implement Linkerd service mesh patterns for lightweight, security-focused service mesh deployments. Use when setting up Linkerd, configuring traffic policies, or implementing zero-trust networking with minimal overhead.

    Also in 5 projects that use this

  • berntpoppSkill

    dependency-cve-sweep

    Use when handling Dependabot PRs, CodeQL/Trivy alerts, or doing a routine dependency/CVE sweep for this repo.

    Also in 12 projects that use this

  • berntpoppSkill

    fleet-standard-adoption

    Use when bringing this repo into compliance with a GeneFoundry fleet standard (tool-naming, response-envelope, container-hardening, versioning, MCP transport, or logging) or closing its tracking issue.

    Also in 12 projects that use this

  • HABUBUSSSkill

    project-guidelines-example

    Also in 16 projects that use this

  • niloykumarbarmanSkill

    agent-sort

    Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.

    Also in 13 projects that use this

  • royaluniondesign-sysSkill

    configure-ecc

    Interactive installer for Everything Claude Code — guides users through selecting and installing skills and rules to user-level or project-level directories, verifies paths, and optionally optimizes installed files.

    Also in 11 projects that use this

  • +2 more from berntpopp

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 167 edited versions we read keep these.

  • 2

    Distinctive

    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:

    Do general quality/readability review (that's a separate concern).

    line 31
  2. Distinctive:

    Trace the data: Follow untrusted input from entry (source) to dangerous operation (sink): query, shell, filesystem, HTTP call, HTML, …

    line 44

“Common” is measured: kept by at least 6 in 10 of the 167 edited versions of this file we could read. “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

Security Review

Find vulnerabilities and explain how they're exploited. Think like an attacker: trust nothing from outside the trust boundary. Report each finding with a severity, location, a concrete exploit scenario, and a remediation.

Severity: 🔴 Critical · 🟠 High · 🟡 Medium · 🟢 Low (rate by exploitability × impact).

Where this skill adds value

Be honest about what this buys you. On a strong model, this skill does not find more bugs than an unaided review — evals showed no detection lift. Its value is consistency and coverage: it works through the full vuln-class checklist so nothing is skipped, rates every finding on one severity scale, and always emits the exploit→impact→fix structure. That reliability matters most for a team standard, a review of record, a large/real codebase, or a weaker model. So: don't just report the one obvious bug — deliver a complete, uniformly rated, well-structured pass.

Scope

This skill DOES:

  • Review for the vulnerability classes in references/vuln-classes.md.
  • For each finding: exploit scenario, impact, and a specific fix.

This skill does NOT:

  • Do general quality/readability review (that's a separate concern).
  • Exploit live systems or write working exploit payloads — describe the attack, don't weaponize it.
  • Silently "fix" code — report findings; the author remediates.

Steps

Surface as TodoWrite items. Checkpoint findings to ./.claude-skills-work/security-review/<target>/findings.md as you go.

  1. Map the attack surface. Identify every point where untrusted data or an attacker interacts with the change: request inputs, params, headers, file uploads, external API responses, deserialized data, auth boundaries, and anything touching secrets or sensitive data.
  2. Trace the data. Follow untrusted input from entry (source) to dangerous operation (sink): query, shell, filesystem, HTTP call, HTML, deserializer. An unsanitized source-to-sink path is a finding.
  3. Check each relevant vuln class against the surface, using references/vuln-classes.md.
  4. Verify the controls, don't assume them. Confirm authz on THIS resource, parameterized queries, output encoding, secret handling — don't assume the framework does it.
  5. Report, most severe first, in the output format below.

High-value checks (see vuln-classes.md for the full set)

  • Injection — SQL/NoSQL/command/LDAP: untrusted input reaching a query or shell without parameterization/escaping.
  • AuthN/AuthZ — missing auth on a route; missing per-resource authorization (IDOR); privilege escalation; trusting client-supplied identity/role.
  • Secrets — hardcoded keys/passwords/tokens; secrets in logs, errors, URLs, or client code.
  • Sensitive data — PII/credentials logged, returned, or stored unencrypted; weak/rolled-your-own crypto; predictable tokens.
  • SSRF / injection of URLs — user-controlled URLs fetched server-side.
  • Deserialization / path traversal / file upload — untrusted data driving object construction, file paths, or executable content.
  • XSS / output encoding — untrusted data rendered without encoding.
  • Config & transport — debug on in prod, permissive CORS, missing TLS, missing rate limits on public endpoints.

Output format

## Security Review: <target>

### 🔴 Critical / 🟠 High / 🟡 Medium / 🟢 Low
1. **<vuln class>** (path:line)
   - Exploit: <how an attacker abuses this, concretely>
   - Impact: <what they gain>
   - Fix: <specific remediation>

### ✅ Controls verified
- <control you checked and confirmed present>

Red flags (stop and correct)

If you notice…Do this
"The framework probably sanitizes this"Verify it; don't assume a control exists
Auth present, but not checked per-resourceConfirm this user may touch THIS object (IDOR)
Reporting a vuln with no exploit pathShow the concrete source→sink; else it's noise
Writing a working exploit/payloadDescribe the attack; don't weaponize
Only checking the classic OWASP listAlso check secrets, logging, SSRF, deserialization
Trusting client-supplied role/id/priceTreat all client input as attacker-controlled

Done when

  • The attack surface of the change is mapped.
  • Each finding has severity, file:line, an exploit scenario, impact, and a fix.
  • Controls you verified are noted; findings ordered most-severe-first.

References (load only when needed)

  • references/vuln-classes.md — vulnerability classes with what to look for, an exploit sketch, and the fix for each.

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.