superreview-setup

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a provided Bash script (scripts/generate.sh) to perform project reconnaissance and automate file generation. These operations are restricted to standard Unix utilities (find, grep, sed, awk) and are scoped to the local repository directory.
  • [DATA_EXFILTRATION]: No unauthorized data exfiltration or external network requests were found. The skill reads local build files (e.g., package.json, pom.xml) and directory structures solely to configure the generated review artifacts.
  • [REMOTE_CODE_EXECUTION]: The skill does not download or execute remote scripts. It uses local templates to create project-specific instructions for the agent.
  • [PROMPT_INJECTION]: The generated agent prompts (such as intent-guard.md) include dedicated sections for 'Non-goals' and 'Checklists' that explicitly define operational boundaries, helping to prevent agent drift and instruction overrides.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 06:23 PM
Security Audit — agent-trust-hub — superreview-setup