init-deep

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of instructional Markdown content. It does not include any Python scripts, shell commands, Node.js files, or other executable binaries.
  • [SAFE]: No malicious patterns related to prompt injection, data exfiltration, or persistence were detected. The skill focuses on organizing project metadata into local documentation files.
  • [EXTERNAL_DOWNLOADS]: While the skill references a GitHub URL for attribution, it does not provide any instructions or automated processes to download, fetch, or execute content from the internet.
  • [PROMPT_INJECTION]: The instructions do not contain commands designed to override system safety filters or reveal the AI's internal system prompt.
  • [DATA_EXFILTRATION]: The skill involves reading a local directory structure to generate documentation. It does not define any network endpoints or transport mechanisms for sending data outside the user's environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 10:20 PM