to-issues
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (
gh issue view) to fetch data from external repositories and system environment to read issue details. - [DYNAMIC_EXECUTION]: The skill instructs the agent to run Python scripts (
recompute_summary.py,check_ac_refs.py) from a local directory. These scripts are used to perform calculations and update board summaries directly on the filesystem. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, such as user stories, acceptance criteria, and GitHub issues, to generate new content and influence downstream tool execution.
- Ingestion points: Data enters the agent context through
gh issue viewcommands and by reading files located indocs/business/anddocs/decisions/. - Boundary markers: The instructions do not specify any delimiters or safety prompts to prevent the agent from following instructions that might be embedded in the processed user stories or PRD files.
- Capability inventory: The skill is capable of writing to the local filesystem (
docs/task-breakdown/), executing Python scripts with write permissions, and creating new GitHub issues with specific labels and assignees. - Sanitization: There is no evidence of sanitization or filtering applied to the external text before it is used to generate card titles, issue bodies, or script inputs.
Audit Metadata