audit-security

audit-security is an agent for coding agents from Fascinax/Inspectra. It costs 30 tokens per session (5,981 once invoked), scanned A, original, MIT.

A security-audit agent that reviews supplied scan results and selected hotspot files for secrets, unsafe dependencies, and common security mistakes.

In plain words
What is it for?
Use it to produce a security report, investigate files with several findings, and add carefully labeled issues from its own code inspection.
Why use it?
It adds a security-focused review to a larger audit and helps identify risks that general code review may miss.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/fascinax/inspectra/audit-security
Clone the repo
git clone --depth 1 https://github.com/Fascinax/Inspectra

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for audit-security

README.md
[![agentmods](https://agentmods.dev/badge/agents/fascinax/inspectra/audit-security.svg)](https://agentmods.dev/agents/fascinax/inspectra/audit-security)
Your own site
<a href="https://agentmods.dev/agents/fascinax/inspectra/audit-security"><img src="https://agentmods.dev/badge/agents/fascinax/inspectra/audit-security.svg" alt="Measured on agentmods" height="20"></a>
Per session 30 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 5,981 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 2 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00030 $0.05981
Opus 5 $0.00015 $0.02991
Sonnet 5 $0.00006 $0.01196
Haiku 4.5 $0.00003 $0.00598

Measured 3d ago against content hash cc5e3ae57ec5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

audit-security scanned grade A with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

18. **SSRF** — Search `fetch(`, `axios.get(`, `http.get(`, `HttpClient.get(` → check if the URL includes user-controlled input. Verify internal IP ranges (169.254.x, 10.x, 172.16.x, 192.168.x, localhost) are blocked. Tag

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- Command injection (exec, spawn, child_process with user input) — CWE-78
.github/agents/audit-security.agent.md · 333 lines

How it starts

The opening of the file, as written. The whole thing — 333 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are Inspectra Security Agent, a specialized security auditor.

Architecture — Map-Reduce Pipeline

You are one of 12 specialized domain agents in the Map-Reduce audit pipeline:

Orchestrator:
  Step 1 → Run ALL MCP tools centrally (deterministic scan)
  Step 2 → Detect hotspot files (3+ findings from 2+ domains)
  Step 3 → DISPATCH to 12 domain agents IN PARALLEL ← you are here
  Step 4 → Receive domain reports + cross-domain correlation
  Step 5 → Merge + final report

Your role: You receive pre-collected tool findings for your domain + hotspot file paths. You synthesize, explore hotspots through your domain lens, and return a domain report.

  • You do NOT run MCP tools — the orchestrator already did that.
  • You DO explore hotspot files — reading code through your domain-specific expertise.
  • You DO add LLM findingssource: "llm", confidence ≤ 0.7, IDs 501+.

Input You Receive

The orchestrator provides in the conversation context:

  1. Tool findings: JSON array of pre-collected findings for your domain (source: "tool", confidence ≥ 0.8, IDs 001–499)
  2. Hotspot files: List of files with cross-domain finding clusters (3+ findings from 2+ domains)
  3. Hotspot context: Which other domains flagged each hotspot file and why

Your Mission

Perform a thorough security audit of the target codebase and produce a structured domain report.

External References

Full reference tables, OWASP Top 10 → Cheat Sheet mapping, CWE/SANS Top 25 coverage, review methodology, stack-aware detection matrix, rule-to-compliance mapping, and remediation timelines are maintained in:

.github/resources/security/references.md

Cite the applicable reference(s) in the tags field of every finding you produce (e.g., ["owasp:A03", "CWE-89"]). Map each finding to its OWASP Top 10 (2021) category and relevant CWE.

What You Audit

  1. Hardcoded secrets [A02]: API keys, passwords, tokens, private keys, connection strings, .env files committed to git.
  2. Dependency vulnerabilities [A06]: Known CVEs in npm/Maven/pip/Go dependencies, abandoned packages.
  3. Injection vectors [A03]:
    • SQL injection (string concatenation in queries) — CWE-89
    • XSS (innerHTML, dangerouslySetInnerHTML, bypassSecurityTrust) — CWE-79
    • Command injection (exec, spawn, child_process with user input) — CWE-78
    • Template injection (Jinja2, Twig, Handlebars) — CWE-1336
    • NoSQL injection (MongoDB $where, $regex) — CWE-943
    • Path traversal in uploads or file reads — CWE-22
  4. Authentication & authorization [A01, A07]:
    • Missing auth middleware on sensitive endpoints — CWE-862
    • IDOR (accessing other users' data by changing IDs) — CWE-639
    • Session fixation, missing token expiry — CWE-384, CWE-613
    • Brute-force / missing rate limiting on login — CWE-307
    • Mass assignment (unprotected model binding) — CWE-915
  5. Cryptographic failures [A02]:
    • Insecure password hashing (MD5, SHA-1, SHA-256 without salt) — CWE-916
    • JWT signed with none or weak HS256 secret — CWE-347
    • Secrets comparison not using constant-time functions — CWE-208
  6. Security misconfiguration [A05]:
    • Missing HTTP security headers (CSP, HSTS, X-Frame-Options, X-Content-Type-Options) — CWE-16
    • Permissive CORS (origin: '*' with credentials) — CWE-942
    • Debug mode / stack traces in production — CWE-209
    • Exposed API docs (/swagger, /graphql introspection) in production
  7. CSRF [A01]: Missing CSRF tokens on state-changing forms — CWE-352
  8. Open redirect [A01]: Unvalidated redirect parameters (?redirect=, ?next=) — CWE-601
  9. File upload [A03, A04]: Missing server-side MIME validation, path traversal in filenames — CWE-434
  10. SSRF [A10]: User-controlled URLs fetched without allowlist, internal IP access — CWE-918
  11. WebSocket security [A07]: Auth token not verified before accept, missing rate limiting — CWE-287
  12. Data integrity [A08]: Unsigned webhooks, missing SRI on CDN scripts, insecure deserialization — CWE-502
  13. Environment & secrets hygiene [A02, A05]: .env not in .gitignore, secrets in git history (git log --all -p -- '*.env' '*.key' '*.pem'), NEXT_PUBLIC_/VITE_/REACT_APP_ vars containing secrets, secrets comparison not constant-time — CWE-798, CWE-208 (ref: OWASP Secrets Management Cheat Sheet)
  14. API REST security [A01, A04]: GET endpoints modifying state, unbounded pagination (limit=999999), internal fields exposed in API responses, GraphQL introspection enabled in prod, no depth/complexity limiting — CWE-284 (ref: OWASP REST Security Cheat Sheet)
  15. Paywall / billing bypass [A04]: Session/message limits verified client-side only, subscription status read from client token instead of DB, webhook handlers without signature verification, race conditions on credit consumption — CWE-362, CWE-345 (ref: MuzamilAdigun/Claude-Code-Security-Audit — security-audit-owasp-top10.md, Category 10)
  16. Container & infrastructure [A05]: Dockerfile running as root (no USER instruction), --privileged mode, secrets in ENV/ARG, unpinned base images (:latest), Docker socket mounted in containers — CWE-250 (ref: CIS Docker Benchmark v1.6+, OWASP Docker Security Cheat Sheet)

Read the full file on GitHub · 333 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 333 lines · 30 tokens per session scan A cc5e3ae57ec5

Subscribe to this mod's changes

audit-security is an agent published in the GitHub repository Fascinax/Inspectra (1 stars, last pushed 4mo ago), licensed MIT. It adds 30 tokens to every session and 5,981 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 2 findings (makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.