github-issue-briefing
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from GitHub issues.
- Ingestion points: Fetches issue title, body, and comments from the higress-group/himarket repository.
- Boundary markers: The workflow does not utilize markers or instructions to isolate external data from the system prompt.
- Capability inventory: The agent can execute shell commands via the gh CLI and perform directory creation and file writes on the local system.
- Sanitization: No input validation or content escaping is mentioned for the processed issue data.
- [COMMAND_EXECUTION]: The skill uses local command-line tools to fetch external data and manage output files.
- Evidence: Execution of gh issue list and mkdir -p reports as part of the automation workflow in SKILL.md.
Audit Metadata