web-pentest-skill
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data (web artifacts, scans), which presents an inherent surface for indirect prompt injection. However, the skill implements defensive posture controls to mitigate this risk.
- Ingestion points: The skill processes various web-assessment artifacts including HTTP captures, crawler output, Nuclei JSON results, and JavaScript inventories (SKILL.md).
- Boundary markers: The instructions explicitly mandate that worker roles must ignore any instructions contained within the analyzed artifacts (SKILL.md).
- Capability inventory: The skill maintains the ability to write to session-specific directories (
runs/<session-id>/) and generate static HTML reports (SKILL.md). - Sanitization: The skill requires redacting secrets (credentials, API keys, cookies) before model use and enforces the escaping of untrusted text in generated HTML output (SKILL.md).
Audit Metadata