i2r-gsd-projection-mode

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a deterministic Markdown projection mode that separates the human-readable 'reading layer' from internal governance data. It enforces strict shape requirements to prevent data pollution.
  • [SAFE]: Implements a 'Gate Check' mechanism (gate.check) designed to explicitly scan for and block 'downstream command leaks' (e.g., /gsd: commands) and 'implementation leaks' (HOW/technical details) in generated artifacts.
  • [SAFE]: The provided reference files, including BAD_PRD.md, contain prohibited patterns solely for educational and validation purposes. These patterns are explicitly documented as blocker defects to harden the system's critic agents against indirect prompt injection.
  • [SAFE]: The assembly process is described as a programmatic pure function (i2r_render.py) that does not rely on LLM invocations for the final document assembly, significantly reducing the risk of runtime instruction overrides or hallucinations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 09:28 AM