cmux-test-bisect
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to execute local scripts and CLI tools to manage the bisection process.
- Evidence: Runs
python3 scripts/ci/package_bisect.pywith various subcommands (start,status,next,cleanup). - Evidence: Executes
gh run view,gh search prs,gh workflow run, andgit pushto interact with the repository and CI environment. - [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze data from external sources, which could potentially contain adversarial content.
- Ingestion points: The agent reads CI logs via
gh run view --log-failed(SKILL.md). - Boundary markers: None mentioned for the ingested log data.
- Capability inventory: The skill has permissions to push to git remotes (
git push), trigger GitHub workflows (gh workflow run), and execute local Python scripts (SKILL.md). - Sanitization: The skill suggests using
grepto filter for specific failure strings, but does not provide robust sanitization for the remaining content.
Audit Metadata