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_output to execute a PowerShell command on Windows for the purpose of retrieving the GITCODE_TOKEN from 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
Risk Level
SAFE
Analyzed
Jun 16, 2026, 05:17 PM
Security Audit — agent-trust-hub — gitcode-pr-audit