code-review

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes several CLI tools to perform its intended functions. It uses git for history and diff analysis, gh for PR management, pnpm for build and type verification, and a custom bd (Beads) tool for task tracking. These commands are standard for a development-focused agent and operate within expected parameters.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) because it ingests and analyzes untrusted data from the repository's source code and git logs.
  • Ingestion points: SKILL.md (Steps 1, 2, and 3) instruct the agent to read file contents and git diffs.
  • Boundary markers: Absent. There are no instructions to the agent to ignore embedded commands within the code being reviewed.
  • Capability inventory: The agent can execute shell commands (git, gh, pnpm, bd) and write local files (SKILL.md Step 5).
  • Sanitization: Absent. The skill does not describe any methods for escaping or filtering the ingested code content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 01:47 PM
Security Audit — agent-trust-hub — code-review