agent-code-review-swarm

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill uses npx to execute the ruv-swarm tool directly from the npm registry. This tool is provided by the author ('ruvnet') and is therefore considered a vendor-owned resource.\n- [COMMAND_EXECUTION]: The skill executes multiple shell commands to interact with the GitHub CLI (gh), including retrieving PR data and diffs, and posting review outcomes to the repository.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8).\n
  • Ingestion points: Reads external, untrusted content from GitHub Pull Requests (titles, bodies, and diffs) using gh pr view and gh pr diff in SKILL.md.\n
  • Boundary markers: None detected. The skill does not use delimiters or explicit instructions to isolate untrusted PR content from the agent's core logic.\n
  • Capability inventory: The skill can perform sensitive GitHub actions such as commenting on PRs, requesting changes, and approving reviews via the Bash tool.\n
  • Sanitization: There is no evidence of data sanitization or validation performed on the PR content before it is processed by the AI agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 04:32 PM