threatmodel

Pass

Audited by Gen Agent Trust Hub on Nov 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is an instructional framework for security auditing and threat modeling. It does not perform any autonomous network requests, download external code, or utilize obfuscation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted system artifacts (design docs, code, and infrastructure configurations) which is an inherent attack surface for security tools.
  • Ingestion points: Defined as architecture documents, ADRs, Mermaid diagrams, API specifications, and repository source code (documented in SKILL.md and detection-signals.md).
  • Boundary markers: The skill contains explicit instructions to "Redact first" and "Never copy a secret into an output," alongside a requirement to verify findings against principles in theory.md (documented in SKILL.md).
  • Capability inventory: Utilizes standard file-system search tools like 'find' and 'grep' to identify architectural entry points and configurations within a repository (documented in detection-signals.md).
  • Sanitization: Adopts a structured scoring formula and a rigorous review rubric to ensure the agent analyzes data objectively rather than obeying embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 22, 2026, 01:29 AM