What can this skill help me with?

It helps design Activepieces flows, generate importable JSON, teach concepts, and debug custom pieces.

Will it expose real IDs or require them?

It uses placeholders ({{CONNECTION_ID}}); you must insert real IDs before deployment.

What if I need a different approach or run into errors?

Consult the reference docs (overview.md, flows-json.md, etc.), follow the protocol, and the skill will guide you through exact patterns for flow generation and troubleshooting.

activepieces-expert

Scanned

npx machina-cli add skill tparrin/my-agent-skills/activepieces-expert --openclaw

Files (1)

SKILL.md

2.9 KB

Activepieces Expert Skill

You are an expert in Activepieces, an open-source, AI-first business automation tool. Your goal is to help users design, build, troubleshoot, and deploy Activepieces flows and pieces.

How to use this skill

This skill uses Progressive Disclosure. Do not guess Activepieces JSON schemas or API details. Always load the specific reference files listed below before providing instructions or emitting code/JSON for those topics. Read the reference file thoroughly and adapt the patterns to the user's specific request.

1. Conceptual & General Knowledge

For core concepts, architecture, and glossary (Agents, flows, Datastore, human-in-loop): See overview.md
For tutorials and distilled lessons: See 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

When a user asks to create an Activepieces flow:

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

Source

git clone https://github.com/tparrin/my-agent-skills/blob/main/activepieces-expert/SKILL.mdView on GitHub

Overview

An expert in Activepieces who helps you design, troubleshoot, and deploy flows and custom pieces. It teaches core concepts, guides flow design, and generates valid importable JSON while avoiding guesswork.

How This Skill Works

It leverages reference docs and follows progressive disclosure. First, it clarifies requirements, then proposes the flow design step-by-step, and finally emits importable JSON using flows-json patterns with placeholders like {{CONNECTION_ID}}. It never guesses JSON schemas and instructs you to load the appropriate reference files.

When to Use It

Design a new Activepieces flow from trigger to actions, including data flow and any required human approvals.
Generate a valid importable flow JSON for deployment when some IDs are not yet known, using placeholders.
Debug a custom piece by tracing data mappings and runtime behavior to identify where things go wrong.
Explain core Activepieces concepts (agents, flows, datastore) to a teammate with guidance from overview.md.
Review or scaffold a SaaS integration piece, ensuring alignment with references/pieces-development.md.

Quick Start

Step 1: Clarify requirements — identify Trigger type, Target systems, Data flow, and any approvals.
Step 2: Propose design — outline the sequence (Step 1 trigger, Step 2 actions, etc.).
Step 3: Generate JSON — emit the importable flow using flows-json.md patterns and placeholders like {{CONNECTION_ID}}.

Best Practices

Never guess JSON schemas; always load the relevant reference files before emitting code or JSON.
Ask clarifying questions about Trigger type, Target systems, Data flow, and human approvals.
Follow canonical JSON patterns from flows-json.md and use accurate {{step_slug.path}} expressions.
Use placeholders like {{CONNECTION_ID}} and confirm IDs before deployment.
Document decisions and rationale to facilitate human-in-the-loop reviews.

Example Use Cases

A lead capture flow triggered by a new CRM contact that sends a welcome email and updates Datastore.
An order placement flow that triggers fulfillment steps and inventory updates via SaaS integrations.
A Slack notification after a task completion, routed through a QA approval step with human review.
A weekly report generation flow that includes a human review stage before publishing.
Debugging a custom TypeScript piece to fix a field mapping error in a SaaS integration.

Frequently Asked Questions

Add this skill to your agents