The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external design data from Figma to generate code artifacts, creating a vulnerability surface for indirect instructions.
Ingestion points: Figma design files, layers, and component metadata accessed via the Anima SDK and API (referenced in SKILL.md and official-docs.md).
Boundary markers: None. The skill does not define specific delimiters or instructions to ignore embedded content within the designs.
Capability inventory: The skill is granted Write, Edit, and Bash capabilities, which are used for code generation and validation steps.
Sanitization: The checklist explicitly mandates ESLint, Prettier, and manual visual/accessibility reviews as quality gates for all generated output.
[EXTERNAL_DOWNLOADS]: The skill references official SDKs and documentation from well-known service providers.
Evidence: References to the Anima SDK on GitHub (github.com/AnimaApp/anima-sdk) and NPM (npmjs.com/package/@animaapp/anima-sdk).
[DATA_EXPOSURE]: Includes a validation script that uses authentication tokens to verify service access.
Evidence: The readiness script in SKILL.md uses FIGMA_TOKEN to authenticate against api.figma.com.
Note: The script targets the official Figma API, a well-known service, and the instructions emphasize storing tokens in managed secret managers rather than source code.

anima-prod-checklist