agentleFS
Luis85Luis85/agentic-workflow/.claude/skills/agentic-security-review/SKILL.md

agentic-security-review

Skill0 starsChanged 6 months agoMITView on GitHub

Run the OWASP-aligned agentic security review path — covers goal hijacking, tool misuse, excessive agency, memory poisoning, secrets exposure, handoff failures, and observability. Opt-in from Stage 9 (Review) or the QA track. Not a new track — no state file or slash command namespace. Triggers "agentic security review", "agent security", "security review for agents", "/agentic-security-review".

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Locate the review doc: Read docs/agentic-security-review.md — it contains the 7 risk categories and checklists.

    line 22
  2. Distinctive:

    Do not claim the system is "secure" or "certified" — use "risk-reduced" or "reviewed for known categories."

    line 63
  3. Distinctive:

    Do not skip the finding record — undocumented risks have no mitigation trail.

    line 64
  4. Distinctive:

    Do not create a new track namespace, state file, or slash command — this is an opt-in skill, not a track.

    line 65
  5. Very specific:

    Do not modify docs/agentic-security-review.md during a review — findings go in the finding record, not the review doc.

    line 66

≈ “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

Agentic Security Review

An opt-in review path for identifying and mitigating risks unique to autonomous, tool-using agents. Invoke from Stage 9 (Review) or the Quality Assurance Track.

Scope: internal risk reduction. Not a certification or complete security guarantee.

When to use

  • Feature adds or changes agent tool use.
  • Stage 9 (Review) for agent-facing workflows.
  • QA track scope includes agentic automation.
  • Post-incident or near-miss involving autonomous agent behavior.

Procedure

  1. Locate the review doc. Read docs/agentic-security-review.md — it contains the 7 risk categories and checklists.
  2. Set up a finding record. Copy templates/agentic-security-findings.md to quality/<review-slug>/agentic-security-findings.md. If no QA review is active, create a minimal quality/<feature-slug>/ folder.
  3. Work through each risk category. For every checklist item in docs/agentic-security-review.md (read-only reference — do not edit the canonical doc):
    • Mark [x] in the findings record if satisfied with evidence.
    • Mark open finding in the findings record if not satisfied.
    • Skip with not-applicable + rationale in the findings record if the category doesn't apply.
  4. Assess severity. For each open finding:
    • critical — exploitable, high-impact, blocks release.
    • high — exploitable or likely, requires mitigation before release.
    • medium — possible exposure, mitigate in next sprint.
    • low — theoretical, accepted or deferred with rationale.
  5. Record mitigations and residual risk. Fill in the finding-record fields: evidence, mitigation, residual risk, follow-up.
  6. Write verdict. Complete the Summary and Verdict sections of the finding record.
  7. Link from the review artifact. In Stage 9's review.md or the QA quality-review.md, add a section:
    ## Agentic security review
    - Finding record: `quality/<slug>/agentic-security-findings.md`
    - Verdict: <complete / incomplete>
    - Blocker findings: <list or "none">

Reporting

On completion, report:

Agentic security review complete.
  Finding record: quality/<slug>/agentic-security-findings.md
  Categories reviewed: 7
  Open findings: N (critical: X, high: Y, medium: Z, low: W)
  Blocker findings: <list or "none">
  Verdict: <complete / incomplete>

Do not

  • Do not claim the system is "secure" or "certified" — use "risk-reduced" or "reviewed for known categories."
  • Do not skip the finding record — undocumented risks have no mitigation trail.
  • Do not create a new track namespace, state file, or slash command — this is an opt-in skill, not a track.
  • Do not modify docs/agentic-security-review.md during a review — findings go in the finding record, not the review doc.

References

  • docs/agentic-security-review.md — full risk categories and checklists
  • templates/agentic-security-findings.md — finding record template
  • docs/quality-assurance-track.md — ISO 9001-aligned QA workflow
  • memory/constitution.md — Article IX (Reversibility)

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.