The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses a tool called WebFetch to retrieve official Snowflake documentation for Cortex AI functions identified in the task context. This is a targeted fetch from a well-known service provider relevant to the skill's purpose.
[COMMAND_EXECUTION]: The skill interacts with Snowflake Cortex AI functions (e.g., AI_COMPLETE, AI_CLASSIFY, AI_SIMILARITY) by identifying them in task descriptions to determine relevance. These are specialized API calls within the Snowflake ecosystem.
[PROMPT_INJECTION]: As a context-loading subagent, the skill modifies the primary agent's prompt by prepending a structured preamble based on external and local data. This creates an indirect prompt injection surface.
Ingestion points: Reads task descriptions from flow.json and various context files stored in the .cocoplus/ directory (e.g., .cocoplus/grove/patterns/, .cocoplus/context/).
Boundary markers: The skill uses a specific header [CocoScout — Relevant Context Loaded] to delimit the injected context within the prompt.
Capability inventory: The skill performs file read and write operations within the .cocoplus/ directory and performs network requests via WebFetch.
Sanitization: The instructions do not specify any sanitization, filtering, or escaping of the ingested content before it is interpolated into the agent's prompt preamble.

cocoscout