The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local utility capable of executing shell commands and scripts through defined targets (e.g., using clip add <name> --stdio <cmd>). This is a core feature of the proxy for managing tool access.
[EXTERNAL_DOWNLOADS]: The skill provides functionality to connect to and fetch data from external remote endpoints, including MCP servers, OpenAPI specifications, and GraphQL/gRPC APIs.
[PROMPT_INJECTION]: As the skill routes and processes data from various external sources (APIs and CLI outputs), it possesses an attack surface for indirect prompt injection where untrusted data could influence agent behavior.
Ingestion points: Data is ingested from external API responses and CLI tool outputs processed by the clip gateway.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within tool outputs are defined in the provided documentation.
Capability inventory: The agent can execute shell commands via stdio or cli targets and perform network operations via REST, GraphQL, SSE, or gRPC targets.
Sanitization: The documentation does not specify sanitization or validation routines for the data received from external targets.

clip