gitcode-pr-audit
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The Python script uses
subprocess.check_outputto execute a PowerShell command on Windows for the purpose of retrieving theGITCODE_TOKENfrom environment variables. This execution is safely implemented using hardcoded command templates and restricted scope parameters ("User" and "Machine"), which prevents arbitrary command injection from user-provided input. - [PROMPT_INJECTION]: The skill processes external data (PR titles, descriptions, and comments) from GitCode. This content is formatted into a Markdown table for the agent's context. While the script performs basic formatting sanitization (escaping pipes), malicious instructions embedded in PR content could potentially influence the agent's behavior (Indirect Prompt Injection).
- Ingestion points: PR data retrieved from the GitCode API in
scripts/pr_audit.py. - Boundary markers: The script escapes Markdown pipe characters to maintain table structure but does not use explicit boundary markers or "ignore instructions" wrappers for the processed PR content.
- Capability inventory: Network requests (GitCode API), local file writing (via
--output), and environment variable retrieval. - Sanitization: Basic character replacement (
|to\|and newlines to spaces) for table stability.
Audit Metadata