github-contribution-period-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/github_contribution_report.pyexecutes the GitHub CLI commandgh auth tokenusingsubprocess.runto retrieve authentication tokens from the user's local environment when other token sources (e.g., environment variables) are unavailable. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data fetched from GitHub, including PR titles and commit messages, which are then interpolated into a Markdown report. This content could potentially contain malicious instructions intended to influence an AI agent that subsequently processes the report.
- Ingestion points: Data is fetched from the GitHub Search API in
scripts/github_contribution_report.py. - Boundary markers: The generated report uses standard Markdown structures (headers, tables, and lists) to organize data.
- Capability inventory: The script includes capabilities for local command execution (
subprocess.run), local file writing, and network communication withapi.github.com. - Sanitization: The
md_escapefunction provides basic escaping for Markdown table delimiters and newlines to maintain report formatting.
Audit Metadata