architecture-decision-record

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input to generate file system outputs, creating a potential injection surface.\n
  • Ingestion points: The skill reads technical-requirements.yaml from the project's requirements directory.\n
  • Boundary markers: Instructions lack explicit delimiters or warnings to treat ingested data as non-executable text.\n
  • Capability inventory: The skill performs directory creation (mkdir) and multiple file write operations.\n
  • Sanitization: No validation or escaping is performed on the extracted decision titles before they are incorporated into file names or ADR body content.\n- [COMMAND_EXECUTION]: The skill executes shell commands to maintain its directory structure.\n
  • Evidence: Step 4 requires the execution of mkdir -p {base_directory}/architecture/adrs to prepare the ADR storage location.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 04:00 PM
Security Audit — agent-trust-hub — architecture-decision-record