implement-spec
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes specifications and ticket data from local files and user input, creating a surface where malicious instructions embedded in these documents could influence the agent's behavior.
- Ingestion points: The agent reads specification documents, docs/agents/issue-tracker.md, and ticket files located in .scratch//tickets/.
- Boundary markers: The instructions do not specify explicit delimiters or warnings to isolate the content of tickets from the agent's execution logic, except for identifying difficulty markers.
- Capability inventory: The skill allows the agent to perform extensive Git operations (branching, worktrees, merging, PR creation), execute local code via tests using the tdd skill, and write files both within and outside the repository.
- Sanitization: There is no evidence of sanitization or validation of the input data processed by the subagents.
- [COMMAND_EXECUTION]: The skill requires the agent to execute various system and shell commands to manage the development lifecycle.
- Evidence: Instructions include creating integration branches, managing multiple worktrees, merging code, and running isolated and shared-state tests. These are legitimate operations for the skill's stated purpose of implementing code from specifications.
Audit Metadata