The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically the Scalekit live catalog and Scalekit MCP, which include tool names, descriptions, and input schemas.
Ingestion points: Data enters the context via the tools.list_tools and search_tools API calls, as well as the documentation page at https://docs.scalekit.com/agentkit/connectors.md.
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for processing these external tool definitions.
Capability inventory: The skill utilizes network operations to interact with Scalekit APIs and performs file-system operations to update .env files.
Sanitization: No explicit sanitization or filtering of the external tool metadata is mentioned before presenting it to the agent or user.
[EXTERNAL_DOWNLOADS]: The skill downloads and installs dependencies including scalekit-sdk-python and @scalekit-sdk/node. These are official SDKs provided by the vendor, Scalekit.
[CREDENTIALS_UNSAFE]: The skill instructs the agent to collect API credentials (SCALEKIT_CLIENT_ID, SCALEKIT_CLIENT_SECRET) from the Scalekit dashboard and store them in a project environment file. This is a standard and recommended practice for managing secrets using environment variables instead of hardcoding them in source code.

discover-connectors