Activepieces Expert Skill

How to use this skill

1. Conceptual & General Knowledge

• For core concepts, architecture, and glossary (Agents, flows, Datastore, human-in-loop): See

  references/overview.md

• For tutorials and distilled lessons: See

  references/tutorial-index.md

2. Flow Design & Execution

• For basic flow design, triggers, and actions: See

  references/flows-basics.md

• For advanced flow concepts (Passing data, versioning, technical limits): See

  references/flows-advanced.md

3. Generating JSON Flows (Import/Export)

• For the canonical JSON schema and expressions to generate importable flows: See

  references/flows-json.md

• CRITICAL: If the user asks for flow JSON, you MUST read

  references/flows-json.md

  before generating the code. Let the user know they will need to insert actual connection IDs.

4. Using and Building Pieces

• For configuring existing pieces (like SaaS apps) and human-in-the-loop patterns: See

  references/pieces-usage.md

• For scaffolding or reviewing custom TypeScript pieces: See

  references/pieces-development.md

5. AI & Agents

• For native AI pieces, building agentic flows, and Datastore: See

  references/ai-and-agents.md

6. Deployment & Troubleshooting

• For common problems (webhooks stopping, connections failing, publishing errors): See

  references/troubleshooting.md

• For cloud vs. self-hosting, Docker, K8s, Railway setups: See

  references/deployment.md

Interaction Protocol for Flow Generation

1. Clarify Requirements: Ask for the Trigger type, Target systems, Data flow, and any human approvals needed. Keep it concise.
2. Propose Design: Outline the flow steps in natural language (Step 1 trigger, Step 2 action, etc.).
3. Generate JSON: Once confirmed, emit the JSON using the patterns in

   flows-json.md

   . Use accurate

   {{step_slug.path}}

   expressions. Include placeholders like

   {{CONNECTION_ID}}

   .