The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the user to connect to an external MCP server at https://rube.app/mcp. This domain is a third-party service and does not belong to the established trusted organizations or well-known cloud services.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources (Amplitude user profiles, event streams, and cohorts) which creates an attack surface for indirect prompt injection.
Ingestion points: Data is ingested via tools such as AMPLITUDE_GET_USER_ACTIVITY, AMPLITUDE_FIND_USER, and AMPLITUDE_LIST_COHORTS.
Boundary markers: The instructions lack explicit delimiters or "ignore instructions" warnings to the agent when handling content retrieved from these external sources.
Capability inventory: The agent has access to state-changing tools including AMPLITUDE_SEND_EVENTS, AMPLITUDE_IDENTIFY, and AMPLITUDE_UPDATE_COHORT_MEMBERSHIP which could be triggered by malicious instructions in the ingested data.
Sanitization: No sanitization or validation logic is specified for data retrieved from the Amplitude API before it is processed by the agent.

amplitude-automation