skills/phrazzld/claude-config/retro/Gen Agent Trust Hub

retro

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the capture of untrusted user input (implementation feedback) and writes it to a persistent local file (.groom/retro.md).\n
  • Ingestion points: User input captured through interactive prompts (e.g., issue numbers, scope changes, blockers) or command arguments.\n
  • Boundary markers: The skill uses Markdown headers (e.g., ## Entry:) but does not implement specific delimiters or 'ignore' instructions to prevent a downstream agent from accidentally executing commands embedded in the feedback.\n
  • Capability inventory: The skill appends data to a local file system path within the repository context.\n
  • Sanitization: No explicit sanitization, escaping, or validation of the user-provided input is performed before writing to the file.\n- [NO_CODE]: This skill does not include any executable scripts or binaries, relying solely on natural language instructions for the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 02:26 PM