to-prd

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) to perform repository management tasks, including gh label list, gh api for fetching milestones, and gh issue create for task tracking. These actions are triggered based on project documentation and user approval.
  • [DYNAMIC_EXECUTION]: The skill executes a Python script (build_ac_index.py) located in a sibling skill's directory. The path is determined by the environment variable or location of the us-ac-formatter skill, which is used to regenerate local documentation indexes.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from decision logs and conversation history to draft user stories and acceptance criteria.
  • Ingestion points: Reads from docs/decisions/ and conversation history buffers.
  • Boundary markers: None explicitly defined to separate instructions from project data in the input files.
  • Capability inventory: Includes the ability to write to the local file system, execute local scripts, and interact with GitHub via the CLI.
  • Sanitization: The skill relies on a human-in-the-loop verification step, showing a full draft to the user for approval before performing any impactful actions like issue creation or repository commits.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 03:31 AM