The Agent Skills Directory

[SAFE]: The skill defines a controlled workflow for documentation lookup using specific Model Context Protocol (MCP) tools. By routing external access through an MCP server rather than generic network operations, it maintains a structured and auditable boundary.
[SAFE]: The inclusion of a "No sensitive data" policy demonstrates high safety awareness. It explicitly instructs the agent to treat user queries as potentially containing secrets and to redact such information before calling the documentation tools, mitigating accidental data exposure.
[PROMPT_INJECTION]: The skill facilitates the ingestion of external data from documentation sources, which creates a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context via the query-docs tool output as described in SKILL.md.
Boundary markers: There are no explicit instructions for using delimiters or boundary markers to isolate the fetched documentation content from the agent's core instructions.
Capability inventory: The skill's primary focus is information retrieval and code generation; no high-risk system capabilities (such as shell execution or file-system modification) are directly triggered by the processing of this external data within the skill.
Sanitization: While the skill mandates sanitization of outgoing data for privacy, it does not specify validation or sanitization protocols for the incoming documentation content.

documentation-lookup