heyeddi-ci-guide
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and act upon data from external sources, such as pull request review comments and GitHub Check results. This introduces an indirect prompt injection surface.\n
- Ingestion points: Handling of failing GitHub Checks and PR review feedback as described in SKILL.md and context/EXAMPLES.md.\n
- Boundary markers: The skill suggests using context markers like and (reference/commands.md) to delineate instructions.\n
- Capability inventory: The package includes a command execution utility in scripts/_skill_cli.py using subprocess.run.\n
- Sanitization: There are no explicit instructions for the agent to sanitize or validate the external data before processing it.\n- [COMMAND_EXECUTION]: The utility script scripts/_skill_cli.py defines a run_command function that wraps subprocess.run. While the skill is described as documentation-only with no active tools in the manifest, the inclusion of this executable helper increases the functional footprint of the skill package.
Audit Metadata