security-pen-testing
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the analysis of untrusted external data, which creates a potential vector for indirect prompt injection attacks where malicious instructions hidden in processed files could influence the agent's behavior.
- Ingestion points: The
vulnerability_scanner.pyscript reads and analyzes source code files (line 120),dependency_auditor.pyreads package manifests (line 173), andpentest_report_generator.pyprocesses findings from JSON files (line 320). - Boundary markers: The instructions do not specify explicit boundary markers or "ignore instructions" delimiters for the agent to use when interpreting the content of the files it scans.
- Capability inventory: The skill scripts have the capability to read from the filesystem, perform regex-based analysis, and write report files to the filesystem.
- Sanitization: The scripts perform basic text processing and regex matching without explicit sanitization or filtering for natural language instructions that might be embedded in code comments or data fields.
Audit Metadata