nika-authoring

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as an instructional guide for a specific Domain Specific Language (DSL). It explicitly advocates for a 'default-deny' security posture by requiring authors to define a permits: block for all filesystem, network, and tool access.
  • [SAFE]: The document provides clear guidance on managing sensitive information using a dedicated secrets: authority, discouraging the use of environment variables or hardcoded credentials.
  • [SAFE]: Although the skill describes advanced features like the lift: block (which can bypass certain static checks for 'taint' or 'data-as-code'), these are documented as sanctioned, reviewable features requiring human-readable justifications rather than covert exploitation techniques.
  • [SAFE]: External domain references (e.g., github.com, rust-lang.org) are used strictly as examples for teaching users how to configure network access boundaries.
  • [SAFE]: The skill includes instructions to use automated checking tools (nika check) to identify and repair security vulnerabilities (such as NIKA-SEC-004) before workflows are executed, which acts as a built-in safety guardrail.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 04:38 PM
Security Audit — agent-trust-hub — nika-authoring