concept-build

Pass

Audited by Gen Agent Trust Hub on Nov 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a reference guide for modular software development. It does not perform network operations (other than referencing well-known technical and academic sources), does not handle sensitive data, and does not execute arbitrary code. All external references target reputable academic domains (arxiv.org, mit.edu), official documentation (gradle.org, rust-lang.org, spring.io), or the author's own repository.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied specifications (such as CONCEPT.md or SYNCS.md) to generate code implementations. While this creates a theoretical surface for indirect prompt injection via malicious design documents, the skill's instructions emphasize strict adherence to architectural boundaries, type safety, and contract preservation, which act as inherent mitigations. This behavior is fundamental to the skill's primary purpose of code generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 1, 2026, 11:23 AM