yds-report-to-issues
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (
gh) andjqto perform repository management tasks. Specifically, it lists existing issues, manages labels, and creates new issues. These operations are restricted to the repository context and are triggered only after explicit user confirmation. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from markdown files located in
docs/evaluation/anddocs/security-audit/. - Ingestion points: Files in
docs/evaluation/*.mdanddocs/security-audit/*.md. - Boundary markers: The workflow includes Phase 3 ("Present Tasks for Selection") which acts as a mandatory human-in-the-loop checkpoint, preventing the agent from automatically acting on malicious instructions embedded in the reports.
- Capability inventory: The skill can execute shell commands via the
ghCLI to create issues and labels (SKILL.md, Phase 5 & 6). - Sanitization: While no explicit regex-based sanitization is mentioned, the skill requires the user to manually select and confirm the content of each task before the
gh issue createcommand is executed, providing a strong defense against automated exploitation.
Audit Metadata