fusion-ai-builder

Plugin fusion-ai-builder from marketplace public-agent-plugins in marat-ak/PublicAgentPlugins. Commands read from the repo's manifest — 2 days ago.

Skills (18)

• agents-topics-policies

  Use when the request involves an AGENT node, a reusable agent artifact (.agent), topics/guardrails/instructions (.topic), a policy store or policy node (.policy), or a reusable function template (.function) — including deciding AGENT node vs LLM node and private vs reusable agents.

• aistudio

  Use for creating or modifying AI Studio app and workflow project resources, including agentic apps, workflows, business objects/BO nodes, policies, tools, connectors, approvals, document schemas, and function templates. Route requests, load the relevant bundled prompt references, and use the bundled

• app-authoring

  Use when the user wants an Agentic App (.app), an app panel/agent/action/communication/template, or a workflow that must power an app (InitDisplay, Query, Summary, InitActions, InvokeAction, app-compatible, Agentic App) — including converting an existing chat workflow into an app-backed one.

• approvals

  Use when the request involves an approval process (.approval) — its payload attributes, rule levels, approvers, notification channels or email account — when a HUMAN approval node must point at an approval process, or for approval lookups (users, email accounts, manager types, representative types,

• business-objects-and-tools

  Use when a workflow needs Fusion data (which business object, which function, which parameters), when wiring a BO_FUNCTION node, when creating a Business Object (.bo) or a Tool (.tool — business-object, deep-link, document, email, external-rest, mcp, connector), or when a request says "BO tool", "de

• chat-debugger

  Use when the user wants to try, chat with or debug the workflow being built THEMSELVES ("let's debug it", "let me try it", "open the debugger"), asks what happened in the debugger panel or why a run there failed, stopped or waited, wants the panel reset, its conversation exported, breakpoints set th

• connectors

  Use when the request involves a connector definition (generated from an OpenAPI or MCP spec), a connector instance (a configured connector), a connector tool, or Knowledge Management connector discovery (list-km-connectors, list-km-connector-tools).

• deeplinks

  Use when the request involves a deeplink artifact (.dl — a Fusion page URL with tokens), a deep link tool that opens one, or deep link discovery on the pod (list-deep-links, search-deeplink-tools).

• document-schemas

  Use when the request involves a document schema (.documentSchema), a Draft-07 JSON Schema generated from one source document — creating it, attaching or replacing the source document, generating or editing the schema, validating, customizing a seeded schema, or saving it.

• gnimsys-chat-widgets

  Use when a chat workflow talks to the user through OUR chat component (`xxx-gnimsys-ai-chat` — the user says so, never assumed) and a turn needs a selection or a decision — a select with inline options or a server-searched datasource (Business Objects API by default, Fusion REST or a BI Publisher re

• import-export

  Use when delivering any artifact as a file (fai_export of a workflow, app, agent, tool, BO, deeplink, topic, policy, policy template, function, document schema, approval or connector; presenting the file with ui_present_file; version stamp; Studio import steps), when the user uploads an export or ar

• instance-import-export

  Use when the user asks to pull an artifact and its dependencies from a Fusion pod into the workspace, to save a workspace artifact and its dependencies to the pod as DRAFTs, to move artifacts between instances, or asks what fetch or save does to local and server state.

• pod-operations

  Use when the user explicitly asks to connect to a Fusion pod or work against one (fetch or save DRAFTs, find a pod workflow by name or code, search pod business objects / deep links / connectors, remote lists, runs and tests, server-side generation), or when a fai tool answers pod-required, auth-req

• policy-templates

  Use when the request involves a policy template (.policyTemplate) — the prompt and function contracts a template-based policy is generated from — creating, editing, validating, listing, fetching or saving one, or resolving a policy's policyTemplateCode.

• run-and-debug

  Use when the user wants to run or test a workflow or app on the pod, see what a run did (executed nodes, node inputs and outputs), stop a run before a node (a breakpoint), run a batch of conversation scenarios from an uploaded file (scenario runs), rerun, override or pin node config/outputs for debu

• validation-and-debugging

  Use when fai_validate or a validate-* command reports errors or warnings (any of the 14 artifact types), when an imported or fetched artifact fails validation, or when you need to understand an existing workflow's structure and real execution path. Running and the debugger itself: run-and-debug; the

• widgets-and-forms

  Use when a workflow node must render rich UI in an Agentic App — tables, cards, charts, clickable lists, an editable form with select fields, hidden state, a submit round trip (oraInfoDisplay, recordWidget, multiRecordWidget, oraFormSubmit, App Experience layouts, the llmGUI dual path every widget s

• workflow-authoring

  Use when creating or modifying any Studio workflow (.wf) — adding, editing, wiring or deleting nodes, writing prompts or expressions, choosing node types, guarding data dependencies, or fixing control-flow/schema errors — whether the workflow is standalone or backs an app.

Hooks (1)

• PreCompact

  ${CLAUDE_PLUGIN_ROOT}/hooks/pre-compact.sh

Similar plugins