The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests component metadata from the external Figma API, which represents a potential attack surface if the design source is compromised.
Ingestion points: Figma component data is retrieved via the scanFigmaComponents function in src/pipeline/figma-scanner.ts.
Boundary markers: The batch-generator.ts script uses path.resolve and a prefix check to ensure generated files are restricted to the authorized output directory.
Capability inventory: The skill utilizes fs.writeFileSync for file operations and uses the restricted Bash(npm:*) tool for package tasks.
Sanitization: Employs explicit path validation to mitigate directory traversal risks and advises manual inspection of generated code diffs as a final security gate.
[EXTERNAL_DOWNLOADS]: Communicates with well-known services, including Figma (api.figma.com) and the Anima SDK repository, which is standard for its intended design-to-code functionality.

anima-core-workflow-a