spec

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE] (SAFE): No malicious code or behavior detected. The skill's logic is confined to information gathering and documentation generation.
  • [Indirect Prompt Injection] (SAFE): The skill contains a vulnerability surface for indirect prompt injection as it processes untrusted user input ($1) into a document. 1. Ingestion points: User-provided functional descriptions in the $1 argument. 2. Boundary markers: Absent. 3. Capability inventory: The skill writes a SPEC.md file to the project root. 4. Sanitization: No sanitization is performed on user input, but the risk is assessed as negligible because the output is static markdown intended for documentation and does not invoke executable commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:48 PM