The Agent Skills Directory

[SAFE]: The skill integrates with Amplitude, a well-known analytics service, via a Model Context Protocol (MCP) server. This behavior is documented and aligns with the skill's primary purpose of discovering and instrumenting analytics for Sentry's user interface. All external references trace back to the vendor's (getsentry) established development ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external tool outputs (Amplitude query results) to inform its actions. However, it incorporates multiple safeguards, including mandatory human-in-the-loop checkpoints via AskUserQuestion before any code modifications are performed.
Ingestion points: Event discovery and data querying tools (mcp__claude_ai_Amplitude__search, query_dataset).
Boundary markers: The skill explicitly instructs the agent to ask the user for confirmation before proceeding to instrumentation if an event is not found.
Capability inventory: The skill allows for searching the local codebase and modifying event definition files.
Sanitization: While structured data from the MCP is expected, the requirement for user confirmation before implementation serves as the primary mitigation against malicious data influence.

analytics