piece-output-schema

Generates outputSchema definitions for Activepieces piece actions and triggers from captured live output.

24.2k|4.1k|Updated Dec 3, 2022
One-click install
npx skills add https://github.com/activepieces/activepieces --skill piece-output-schema
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: piece-output-schema
Source: https://github.com/activepieces/activepieces/tree/main/.agents/skills/piece-output-schema
Command: npx skills add https://github.com/activepieces/activepieces --skill piece-output-schema

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Activepieces steps return raw JSON that renders as an unstructured dump in the flow builder's data selector, making it hard for users and AI/MCP consumers to find the fields that matter. This Skill turns that raw output into a curated, labelled, typed tree by authoring outputSchema definitions grounded in real captured data.

Core Features & Use Cases

  • Schema authoring from real output: Captures actual step output via the builder Test Step or the test-step API, then writes curated schemas in output-schemas.ts with labels, formats, and relative paths.
  • Full workflow coverage: Scopes which actions/triggers need schemas, handles top-level arrays, dynamic-key maps, shared field-sets, and wires schemas into actions, triggers, or trigger factory maps.
  • Validation and release hygiene: Verifies every path resolves against captured JSON, bumps piece patch versions, rebuilds dev pieces, and runs lint checks.
  • Use Case: A contributor adds a new action to a community piece and wants its output to render as a friendly tree in the builder; the Skill runs the step against a live connection, captures the JSON, and authors a validated schema.

Quick Start

Ask the AI to generate outputSchema for a specific Activepieces piece, providing the piece name and a working connection to use for capturing real step output.

Frequently Asked Questions about piece-output-schema

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add an outputSchema to an Activepieces piece action?

Create or edit output-schemas.ts in the piece's src/lib directory, define an OutputSchema with curated fields, then add outputSchema to the createAction object. Capture real step output first so every field path resolves against the actual return value.

How do I capture real output from an Activepieces step?

Run the step via the builder's Test Step against a live connection, which is the preferred method since the engine refreshes OAuth tokens. Alternatively use the POST /v1/sample-data/test-step API and poll for the output once the queued run completes.

Which actions and triggers should get an outputSchema?

Create, update, get, list, search, and find actions plus stable webhook or polling triggers should get schemas. Skip delete actions returning empty bodies, generic custom_api_call passthroughs, and polymorphic triggers whose payloads vary structurally.

Why does my outputSchema show dead or missing fields in the builder?

Dead fields usually come from absolute child paths: children and listItems paths must be relative to the parent object or array item, not the root. Also confirm the piece version was bumped and rebuilt so the builder serves fresh metadata.

Can I author a list schema from an empty array response?

No. If a list or search returns an empty payload, first run the corresponding create or write action to seed data, then re-run the read with the new record's id. Authoring from an empty array produces schemas with unverifiable paths.