The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Node.js script scripts/validate-vision.js to validate and migrate product vision data stored in vision.json. This is an expected operational step for ensuring data integrity within the project context.
[PROMPT_INJECTION]: The skill features a surface for indirect prompt injection by reading user-generated JSON and Markdown files to facilitate the generation of subsequent project documents.
Ingestion points: vision.json, docs/product-vision.md, and docs/prd.md.
Boundary markers: Not explicitly defined in the intake instructions.
Capability inventory: The skill utilizes file system write access to the docs/ directory and executes Node.js commands.
Sanitization: While schema validation is performed by a local script, the instructions do not specify sanitization for natural language instructions embedded within the data.

plaid-plan