The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute Node.js commands using the node -e flag to probe API endpoints and inspect data shapes. This involves running dynamic module code directly in the shell.
[DYNAMIC_EXECUTION]: The suggested debugging probe uses dynamic import statements and runtime environment modification (globalThis.window = {}) to execute logic outside the standard application lifecycle.
[DATA_EXFILTRATION]: The skill guides the agent to read potentially sensitive configuration from .env and ~/.canvasrc (a hidden file in the user's home directory) to resolve CANVAS_SITE_URL. This access to home-directory configuration presents a risk of exposing local development credentials.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests data from external Drupal JSON:API endpoints using the JsonApiClient and useSWR hooks in SKILL.md.
Boundary markers: None implemented for the API data itself, though the skill mandates a manual verification process for request results.
Capability inventory: The skill performs network operations via fetch and JsonApiClient, and local shell execution via the Node.js probe pattern.
Sanitization: The skill focuses on verifying field presence and structure but does not specify programmatic sanitization of the content retrieved from the Drupal API before it is rendered in React components.

canvas-data-fetching