docs-writer

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command to ensure the local '.artifacts' directory is excluded from Git tracking to prevent accidental check-in of generated documentation.\n
  • Evidence: This instruction is present in references/brief.md, references/design.md, references/prd.md, and references/tdd.md.\n
  • Command: grep -qxF '.artifacts' .git/info/exclude 2/dev/null || echo '.artifacts' >> .git/info/exclude\n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted user input during its discovery phase and interpolates it into document templates.\n
  • Ingestion points: User answers during the adaptive interview process defined in references/discovery.md.\n
  • Boundary markers: Templates in references/prd.md, references/design.md, and references/tdd.md lack explicit delimiters or instructions to ignore potential commands embedded in user input.\n
  • Capability inventory: The agent performs file-writing operations to the local .artifacts/docs/ directory.\n
  • Sanitization: No explicit input validation or escaping mechanisms are defined for processing the gathered requirements before drafting.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 07:00 PM