Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/xsourabhsharma/ai-security-audit-pronpx agentmods add skills/xsourabhsharma/ai-security-audit-pro/security-auditWrote 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.
[](https://agentmods.dev/skills/xsourabhsharma/ai-security-audit-pro/security-audit)<a href="https://agentmods.dev/skills/xsourabhsharma/ai-security-audit-pro/security-audit"><img src="https://agentmods.dev/badge/skills/xsourabhsharma/ai-security-audit-pro/security-audit/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/xsourabhsharma/ai-security-audit-pro/security-audit"><img src="https://agentmods.dev/badge/skills/xsourabhsharma/ai-security-audit-pro/security-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00066 | $0.02028 |
| Opus 5 | $0.00033 | $0.01014 |
| Sonnet 5 | $0.00013 | $0.00406 |
| Haiku 4.5 | $0.00007 | $0.00203 |
Grade A, and why
security-audit scanned grade A with 0 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 11d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 192 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Security Audit Pro
Use this skill for defensive security audits only. The user must own the target or have explicit authorization before any active probing, crawler-based DAST, fuzzing, brute-force simulation, or scanner that sends vulnerability payloads.
Core Rule
Be powerful, but evidence-driven and non-destructive:
- Prefer local source, configs, lockfiles, logs, and official docs over guessing.
- Run passive/local checks first.
- Ask before active DAST on a public or third-party target unless the user already gave explicit authorization for that exact target.
- Never print secrets. If a secret-like value is found, report only file, line, type, and a short redacted prefix/suffix.
- Do not exploit, persist access, bypass auth, dump data, run credential attacks, or cause service degradation.
Audit Levels
passive: local code checks plus one fetch of the supplied URL. Safe default.standard: passive checks plus shallow common-exposure checks on authorized targets. The runner requires--authorizedfor URL targets.active: runs installed active scanners such as Nuclei or OWASP ZAP where authorized. The runner requires--authorizedfor URL targets.- Diff scope: local project mode can add
--diff-base <ref>to focus source scans on changed files from a Git baseline.
If the user says "everything", start with passive and standard where authorization is clear, then explain what would need explicit confirmation for active.
Baselines To Apply
Use these frameworks when triaging and reporting:
- OWASP Top 10:2021 for application risk categories.
- OWASP API Security Top 10:2023 for API-specific risk.
- OWASP ASVS 5.0.0 for verification requirements.
- OWASP WSTG for test coverage.
- OWASP Cheat Sheet Series and Secure Headers Project for remediation.
- CWE IDs when a finding maps cleanly.
Workflow
- Scope
- Identify whether the target is a local project path, localhost URL, staging URL, production URL, API spec, or mixed scope.
- Check for app stack, package manager, lockfiles, framework, API routes, auth/session code, and deployment config.
- Confirm authorization before active testing if the target is public or third-party.
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.
- 11d ago First seen · 192 lines · 66 tokens per session scan A 733209e9ff19
security-audit is a skill published in the GitHub repository xsourabhsharma/ai-security-audit-pro (4 stars, last pushed 2mo ago), licensed MIT. It adds 66 tokens to every session and 2,028 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
performing-ai-assisted-vulnerability-discovery
Using LLMs to accelerate vulnerability research and pentest workflows — generating syntax-valid fuzzing seeds and evolving grammars, fine-tuned mutation dictionaries, parallel agent-based proof-of-vulnerability generation, and evidence-driven passive analysis of real HTTP traffic via the Burp MCP server. Covers…
panguard
AI agent security platform — audit skills, scan for threats, and run 24/7 protection with 9,700+ detection rules.
attackgen-tabletop
Turn a MITRE ATT&CK/ATLAS threat group or case study — or a frontier AI agent deployed inside the organisation — into a full incident-response tabletop exercise (MSEL): scenario narrative, kill chain or agent threat scope, timestamped injects, discussion questions, and a detection-coverage scorecard, using the…
testing-mcp-server-security
Testing Model Context Protocol (MCP) servers and the clients that consume them for tool poisoning, prompt injection via tool descriptions/outputs, over-permissioned and local-credential-stealing tools, config/trust bypasses, and unauthenticated RCE during authorized penetration tests of AI agent infrastructure.
subagent-driven-development
Use when executing implementation plans with independent tasks in the current session.
prompt-injection-auditor
Security audit of LLM system prompts, agent instruction files (SKILL.md, AGENTS.md, CLAUDE.md), and agent configurations against prompt injection attacks. Use when the user wants to (1) audit or harden a system prompt or agent instructions against prompt injection, (2) review an agent skill or system prompt for…