adr-critique

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates entirely on local project documentation (markdown files and LikeC4 models) and does not possess capabilities for network communication or external data exfiltration.
  • [COMMAND_EXECUTION]: The skill uses file system discovery via glob patterns (**/*.c4) to locate relevant architectural models. This is a functional requirement for its purpose and is limited to project-specific file types.
  • [DATA_EXFILTRATION]: No exfiltration vectors were identified. The skill does not use any tools or commands associated with remote data transmission (e.g., curl, wget, or fetch).
  • [PROMPT_INJECTION]: The instructions do not contain attempts to override core safety guidelines or bypass agent constraints. The skill enforces a strict output discipline and approval workflow.
  • [REMOTE_CODE_EXECUTION]: The skill does not download or execute external scripts or packages. It relies on internal logic and local documentation for its auditing process.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 03:39 AM
Security Audit — agent-trust-hub — adr-critique