The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides JavaScript templates (e.g., in scenes/fishbone.md, scenes/flywheel.md, and scenes/treemap.md) that the agent is instructed to write to local files (.cjs) and execute using node. This is used to perform geometric and math calculations for complex diagram layouts.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to run @larksuite/whiteboard-cli via npx. As this is a resource owned by the vendor (larksuite), it is considered an intended dependency for the skill's operation.
[COMMAND_EXECUTION]: The skill relies on shell command execution for its core functionality, specifically calling lark-cli for platform interactions and node for running the aforementioned layout scripts.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided data, such as diagram labels, descriptions, and data points, which are then interpolated into script templates and DSL JSON. This creates a surface for indirect prompt injection, as malicious input could attempt to manipulate the generated script or the final whiteboard content.

lark-whiteboard