The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and analyze local UI source files (e.g., .tsx, .html, .css) to generate design scores. This creates a standard injection surface where instructions embedded in the analyzed files (such as in comments or string literals) could potentially influence the agent's behavior during the scoring or subsequent "Gate Mode" fixing process.
Ingestion points: Target source files provided by the user or identified in the project directory (e.g., src/app/Dashboard.tsx).
Boundary markers: Absent. The instructions do not define specific delimiters or clear prompts to ignore instructions that might be contained within the analyzed source code.
Capability inventory: The skill primarily generates text-based reports and scores. However, it references a "Gate Mode" that involves iterative editing and re-scoring, which may involve file-writing capabilities depending on the availability of the referenced /ss-review tool.
Sanitization: Absent. There is no provided mechanism to sanitize or filter the input from the source files before the LLM processes them.

ui-score