The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted content from external websites, which serves as a potential surface for indirect prompt injection where malicious instructions could be embedded in the analyzed data.
Ingestion points: The skill ingests data from palette.json, typography.json, computed-styles.json, breakpoints.json, and rendered HTML excerpts (e.g., homepage.html) provided in the project output directory.
Boundary markers: No explicit delimiters or boundary markers are defined in the instructions to isolate the external content from the agent's instructions.
Capability inventory: The skill is primarily focused on generating a structured JSON foundation and uses an MCP tool (liberate_design_foundation_validate) for schema validation. It does not have access to general shell execution or network requests within the provided prompt context.
Sanitization: No explicit sanitization or filtering of the input web content is described in the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill's documentation (references/theme-tokens.md) specifies the use of Google Fonts (via gstatic.com) for font substitution. This reference involves a well-known and trusted technology service and is integral to the skill's legitimate functionality of replacing commercial fonts with open-source alternatives.

design-foundations