recording-pentest-engagement

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/record_engagement.py invokes the gpg utility using subprocess.run to sign manifests. The implementation uses a list of arguments, which is a secure coding practice that prevents shell injection attacks. This usage is consistent with the skill's allowed-tools policy.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes findings data which serves as a vulnerability surface for indirect prompt injection. Ingestion points: JSON files within the findings/ directory are read and parsed by scripts/record_engagement.py. Boundary markers: The script implements logic to detect and flag evidence paths that point outside the engagement directory tree, ensuring archive integrity. Capability inventory: File system read/write, cryptographic hashing, archiving via the tarfile module, and external signing via GPG. Sanitization: The script performs data structure validation on parsed JSON finding records and validates path strings before processing evidence fields.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:48 AM
Security Audit — agent-trust-hub — recording-pentest-engagement