code-review-parallel

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes official command-line interfaces (gh, glab, az) to interact with repository platforms. These tools are used for their intended functions, such as fetching pull request metadata and posting review comments.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves code diffs and repository configuration files (CLAUDE.md) from remote hosting providers. These sources (GitHub, GitLab, and Azure DevOps) are well-known technology services.
  • [PROMPT_INJECTION]: The skill processes untrusted content from external repositories, creating a surface for indirect prompt injection.
  • Ingestion points: Repository diffs and CLAUDE.md files are fetched in Step 4 and passed to sub-agents.
  • Boundary markers: The review prompt in Step 5 uses clear delimiters (e.g., --- DIFF ---, --- CLAUDE.md ---) to separate external content from instructions.
  • Capability inventory: The skill executes shell commands for platform interaction and performs file system writes to store review results locally.
  • Sanitization: Content is wrapped in delimiters to prevent unintended instruction execution, though no explicit filtering of the external text is performed.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 07:04 AM
Security Audit — agent-trust-hub — code-review-parallel