implement-to-pr
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads GitHub issue content and comments which can contain malicious instructions intended to manipulate the agent.\n
- Ingestion points: The skill uses
gh issue view <number> --commentsto read instructions from GitHub issues inSKILL.md.\n - Boundary markers: No delimiters or safety warnings are specified to distinguish issue content from system instructions.\n
- Capability inventory: The agent can modify files, execute build/test commands, and interact with GitHub via the
ghCLI.\n - Sanitization: No sanitization of the issue or comment content is performed before processing.\n- [COMMAND_EXECUTION]: The skill is instructed to run arbitrary commands defined within the repository configuration.\n
- Evidence: Step 4 in
SKILL.mdrequires the agent to "run the repository's full checks (formatting, linting, type checks, tests, and build)", which triggers the execution of scripts defined in the project's build system.\n- [DYNAMIC_EXECUTION]: The skill performs runtime execution of code found in the repository to verify the changes.\n - Evidence: The verification process requires running the full test suite and build pipeline of the target repository as described in
SKILL.md.
Audit Metadata