khe-design-tokens
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to follow an agent.instructions field in design files. This creates a surface for indirect prompt injection if the design files are sourced from an untrusted third party.\n- Ingestion points: The agent parses .design and DESIGN.md files for design tokens and instructional metadata.\n- Boundary markers: The skill does not employ specific delimiters to isolate instruction-bearing fields from design data.\n- Capability inventory: The agent can modify files in the current repository and execute the provided local Node.js script.\n- Sanitization: No sanitization is performed on instructions extracted from design contracts.\n- [COMMAND_EXECUTION]: The skill executes a local Node.js script (scripts/token-check.mjs) to verify token usage and recommends the use of trusted external tools for linting design files.\n- Evidence: The skill instructions frequently reference the execution of the token-check.mjs script to validate the work and ensure a green gate pass.
Audit Metadata