skills/nweii/agent-stuff/spec-shaping/Gen Agent Trust Hub

spec-shaping

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • NO_CODE (SAFE): The skill contains no executable scripts, binaries, or package manifests. It relies entirely on instructions for the agent's internal behavior.
  • Prompt Injection (SAFE): No override patterns, safety bypasses, or system prompt extraction attempts were detected.
  • Data Exposure & Exfiltration (SAFE): The skill searches for and reads local project documentation and configuration files (e.g., agents.md) for context but has no mechanism for external network transmission or credential theft.
  • Indirect Prompt Injection (SAFE):
  • Ingestion points: Processes project plans and specification files provided by the user.
  • Boundary markers: No explicit delimiters are specified for ingested content.
  • Capability inventory: Capability is limited to searching, reading, and writing markdown documentation.
  • Sanitization: No sanitization is performed on external content, but the lack of high-risk capabilities (like shell execution) makes the surface non-exploitable.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:45 PM