yeet
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill automates GitHub workflows by executing
ghandgitCLI commands. It uses local helper scripts (yeet-context.sh,issue-form.py) to fetch repository metadata and process YAML issue forms. These executions are well-scoped to the skill's primary purpose and use secure patterns like quoted HEREDOCs to prevent command injection. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub (issue titles, bodies, and comments) which could contain malicious instructions. This risk is acknowledged and mitigated in the instructions through a mandatory 'External-disclosure Review' step and the use of secure delimiters during content generation.
Audit Metadata