The Agent Skills Directory

[SAFE]: The skill operates as a specialized design specification tool with no malicious patterns detected. Its use of the file_write capability is appropriate for generating the intended DESIGN.md and brief-preview.html files, and it does not perform network operations or access sensitive system paths.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted natural language data which could potentially contain malicious instructions.
Ingestion points: The brief input parameter defined in the frontmatter of SKILL.md.
Boundary markers: Not explicitly defined in the prompts.
Capability inventory: The skill is limited to file_write operations for documentation generation.
Sanitization: The risk is mitigated by a strict symbolic-to-concrete token resolution process (Section 2.1) and a validation step that forces the agent to prompt for clarification if the input does not match a closed vocabulary of valid design tokens.

design-brief