fixing-flaky-tests
Pass
Audited by Gen Agent Trust Hub on Jun 24, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local test runners and build tools, including pnpm jest, pytest, and the vendor-specific hogli utility, to reproduce failures and validate fixes.
- [EXTERNAL_DOWNLOADS]: The skill fetches history and logs from the official GitHub Actions API for the PostHog repository using the gh command-line tool.
- [PROMPT_INJECTION]: The skill involves reading and analyzing external content from CI logs. This constitutes an indirect prompt injection surface as logs may contain untrusted data from the code under test, but this is inherent to the skill's purpose of debugging. Ingestion points: Raw job logs retrieved from GitHub Actions via gh api. Boundary markers: Absent. Capability inventory: Execution of local testing tools and GitHub API interactions. Sanitization: No specific sanitization mentioned.
Audit Metadata