ci-triage
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected. The agent processes untrusted data (CI logs and workflow metadata) from GitHub to generate issue titles and bodies. This could allow malicious content in CI logs to influence the agent.
- Ingestion points: CI log excerpts and job metadata referenced in SKILL.md.
- Boundary markers: No explicit boundary delimiters are used to isolate untrusted logs within the issue template.
- Capability inventory: The skill has the ability to read commits and file issues via GitHub tools.
- Sanitization: No sanitization of ingested content is specified.
- [COMMAND_EXECUTION]: The skill uses GitHub MCP tools to fetch commits and list issues. This is necessary for its functionality and aligned with its stated purpose of CI triage.
Audit Metadata