krt-ci-questor
Pass
Audited by Gen Agent Trust Hub on May 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes official platform CLIs such as
gh(GitHub),glab(GitLab), andcirclecito retrieve job logs, annotations, and metadata. - [EXTERNAL_DOWNLOADS]: References external documentation and research literature from well-known and trusted domains including
docs.github.com,docs.gitlab.com,circleci.com, andarxiv.org. - [SAFE]: The skill incorporates explicit safety protocols in
references/safety.mdandSKILL.mdto ensure secrets, tokens, and credentials are never printed or exposed during the investigation process. - [SAFE]: Instructions strictly forbid the agent from mutating remote CI state (e.g., rerunning jobs or changing configurations) without explicit user authorization.
- [PROMPT_INJECTION]: The skill processes untrusted external data in the form of CI logs and repository files, creating a surface for indirect prompt injection; however, the structured playbook and focus on specific failure signatures mitigate the risk of the agent obeying embedded commands.
Audit Metadata