specsfy-data-discovery

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted project data and user input to generate documentation.
  • Ingestion points: Project files such as MVP.md, PROJECT.md, Inbox, and user-provided conversation inputs.
  • Boundary markers: The skill contains explicit instructions to separate confirmed data from observations and has a 'Limits' section that forbids filling gaps by inference.
  • Capability inventory: File system write operations are performed via the included registrar_dados_conversados.mjs script.
  • Sanitization: The recording script sanitizes input strings by escaping markdown pipe characters (|) and removing newlines to ensure the integrity of the generated markdown table.
  • [COMMAND_EXECUTION]: The skill utilizes a local Node.js script to automate the recording of discovered data.
  • The script (registrar_dados_conversados.mjs) is provided as part of the skill and relies solely on Node.js built-in modules (node:fs, node:path, node:promises).
  • No external network requests or third-party package dependencies are present in the script code.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 05:37 PM
Security Audit — agent-trust-hub — specsfy-data-discovery