The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The script scripts/inject-brand-context.cjs processes content from docs/brand-guidelines.md to generate context for the agent's system prompt.
Ingestion points: docs/brand-guidelines.md (via scripts/inject-brand-context.cjs).
Boundary markers: None; the output uses plain text headers such as BRAND CONTEXT: without explicit instructions for the agent to ignore embedded commands.
Capability inventory: The skill uses execFileSync in scripts/sync-brand-to-tokens.cjs and performs file system writes to sync design tokens.
Sanitization: None; content is extracted using regular expressions and interpolated directly into the prompt fragment.
[DYNAMIC_EXECUTION]: The script scripts/sync-brand-to-tokens.cjs uses execFileSync to execute a local node script at .claude/skills/design-system/scripts/generate-tokens.cjs. This is an internal dependency call within the expected directory structure of the skill ecosystem.
[EXTERNAL_DOWNLOADS]: The documentation and starter templates reference Google Fonts (fonts.googleapis.com) for CSS implementation. This reference to a well-known service is standard for typography implementation and is considered safe.

brand