The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands using the bun runtime, specifically bun install for dependency management, bun extract for data retrieval, and bun run typecheck:main for validation within the tools/figma-v3-migration directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes component metadata fetched from the Figma API, which is then stored in .d.ts declaration files for the agent to read. This creates an ingestion point for untrusted external data that could potentially contain malicious instructions intended to influence the agent's code generation.
Ingestion points: Generated metadata files located in src/main/data/__generated__/v3-component-sets/.
Boundary markers: No specific delimiters or instructions to ignore embedded content are provided in the guide.
Capability inventory: The skill possesses shell execution and file-write capabilities across the migration toolset.
Sanitization: The instructions do not specify any validation or sanitization steps for the data fetched from the Figma API.
[EXTERNAL_DOWNLOADS]: The workflow involves fetching component metadata from the Figma API using the bun extract command. Figma is a well-known service, and while this is a standard operational requirement, it represents a data ingestion point from a remote source.
[OBFUSCATION]: The action-sheet.ts code example in the guide contains a backspace control character (\b) within a string literal ("Layout:\bText Only"). While likely a typographical error, such characters can be used to bypass simple text-based security filters or validation logic.

seed-dev-figma-v3-migration-plugin