Plug AI Hub
AI-generatedSummary
Emit the configured AI agent system prompt, maximum tool calls per turn, and forbidden capability names for use in Plug MCP workflows.
Inputs
- identity — Defines the assistant's identity and introduction.
- scope — Specifies the business scope and boundaries for the assistant.
- maxToolCallsPerTurn — Maximum number of tool calls allowed per user message, integer between 1 and 20.
- sensitiveDataRules — Text rules for handling sensitive customer and credential data.
- operationalLimits — Textual operational guardrails enforced at the prompt level.
- forbiddenCapabilityNamesJson — JSON array of capability names that must never be exposed to the agent.
Output shape
A single item containing the properties: systemPrompt (string), maxToolCallsPerTurn (number), forbiddenCapabilityNames (string array), and wiring object with the same properties formatted for Plug MCP Server.
The output does not perform any capability execution but only provides the configured AI assistant system prompt and related parameters for wiring into AI agent system messages and Plug MCP Server configurations. Errors during execution will be outputted if continueOnFail is enabled.