what-next
Pass
Audited by Gen Agent Trust Hub on Nov 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub issue bodies and comments which could contain instructions meant to deceive the agent.
- Ingestion points: SKILL.md uses
gh api graphqlandgh issue viewto retrieve titles, bodies, and comments from the issue tracker. - Boundary markers: The instructions do not define delimiters or specific safety prompts to isolate untrusted data from the agent's logic.
- Capability inventory: The skill uses
ghandnodeto query repository state and recommends shell commands to the user. - Sanitization: No sanitization or filtering of the retrieved text is performed before it is used for classification and ranking.
- [COMMAND_EXECUTION]: The skill executes shell commands based on configuration found in the repository's documentation.
- Evidence: SKILL.md executes
node .agents/skills/handoff/handoff-map.mjsusing keys derived from local settings. - [DYNAMIC_EXECUTION]: The skill follows "discover" instructions found in settings files, which involves executing arbitrary shell commands defined as strings in the repository's documentation.
- Evidence: settings/README.md and SKILL.md define a
discover: <how>pattern where the agent is instructed to run the shell command provided in the<how>placeholder to find configuration values at runtime.
Audit Metadata