Audit candidate: 0xSteph/pentest-ai-agents

[github-actions]

Repo

https://github.com/0xSteph/pentest-ai-agents

Field	Value
Stars	1935
NL Artifacts	55

Next Steps

1. Review the repo — check CONTRIBUTING.md, recent activity, PR history
2. If worthy: add label audit-ready to trigger the audit workflow
3. If not: close this issue

Checklist

• Accepts contributions
• Updated in last 60 days
• Not a fork or mirror
• Has meaningful NL artifacts

[github-actions]

Security Gate: BLOCKED

Critical security patterns detected in 1 file(s).
High-risk patterns in 2 file(s).

This repo will be audited for NL quality but will NOT receive automated PRs.
Manual security review required before any contributions.

[github-actions]

Security Gate: BLOCKED

Critical security patterns detected in 1 file(s).
High-risk patterns in 2 file(s).

This repo will be audited for NL quality but will NOT receive automated PRs.
Manual security review required before any contributions.

[github-actions]

Security Gate: BLOCKED

Critical security patterns detected in 1 file(s).
High-risk patterns in 2 file(s).

This repo will be audited for NL quality but will NOT receive automated PRs.
Manual security review required before any contributions.

[github-actions]

Security Gate: BLOCKED

Critical security patterns detected in 1 file(s).
High-risk patterns in 2 file(s).

This repo will be audited for NL quality but will NOT receive automated PRs.
Manual security review required before any contributions.

[github-actions]

Security Gate: BLOCKED

Critical security patterns detected in 1 file(s).
High-risk patterns in 2 file(s).

This repo will be audited for NL quality but will NOT receive automated PRs.
Manual security review required before any contributions.

[xiaolai]

Audit unsupported (model safety refusal)

The auditor model refused to score this repository — its offensive-security / red-team content (C2, lateral movement, opsec evasion) trips the model's cybersecurity safeguards, so no audit report can be produced. Confirmed reproducible (refused on two separate runs; stop_reason: refusal).

Marking audit-unsupported and closing so the batch-processor stops re-triggering it. No PRs will be opened. Handled going forward by the refusal-detection guard added to auditor-audit.yml + batch-process.py.