The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and modify existing HTML, CSS, and JS files which may contain untrusted data or instructions. * Ingestion points: Reads prototype HTML files from docs/assets/prototype/, AGENTS.md, and external design requirement documents. * Boundary markers: No specific delimiters or boundary markers are used to separate untrusted content from system instructions. * Capability inventory: The skill allows the agent to edit files in the workspace and invoke downstream agent skills ($grill-workflow, $change-scope, $edit-prototype, $spec-with-prototype). * Sanitization: No programmatic sanitization is defined, though the instructions explicitly direct the agent to treat source content as evidence rather than authority and to preserve secrets and role boundaries.

edit-prototype