review-all

Warn

Audited by Gen Agent Trust Hub on Apr 11, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads the @willbooster/agent-skills package from the NPM registry via bunx.
  • [COMMAND_EXECUTION]: The skill mandates a one-hour minimum execution time and explicitly forbids stopping the process ('DO NOT STOP THE COMMAND BEFORE 1 HOUR ELAPSES'). This requirement is highly atypical for static analysis or code review utilities and could be used to facilitate long-running malicious background tasks or ensure persistent execution.
  • [REMOTE_CODE_EXECUTION]: Executes remote code from the author's NPM scope at runtime using bunx.
  • [PROMPT_INJECTION]: The workflow instructs the agent to report valid comments from tool output or respond with a specific phrase, creating a surface for indirect prompt injection where malicious instructions in the analyzed codebase could influence the agent's final report.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 11, 2026, 12:49 AM