Lobster

Automate multi-step workflows with human approval checkpoints and JSON output.

Updated Jun 16, 2026

Search Tags:#automation #workflow #pipeline #json output #approval

Ready to run

npx skills add https://github.com/breakingcircuits1337/agent-skills --skill lobster-breakingcircuits1337

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: Lobster
Source: https://github.com/breakingcircuits1337/agent-skills/tree/main/lobster
Command: npx skills add https://github.com/breakingcircuits1337/agent-skills --skill lobster-breakingcircuits1337

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates multi-step workflows with approval checkpoints, saving time on repeatable automation and ensuring actions require human oversight before execution.

Core Features & Use Cases

Workflows with Approval: Supports automation of workflows with approval steps, ideal for scenarios where actions need human validation.
Structured Output: Always returns JSON enveloped output for consistency and easier integration.
Use Case: Consider automating the process of triaging emails. Fetch recent emails, classify them, and require human approval before any automated action is taken.

Quick Start

Use the lobster skill to run a pipeline: openclaw.invoke lobster run --pipeline 'gog.gmail.search --query "newer_than:1d" --max 20 | email.triage'.

Frequently Asked Questions about Lobster

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

FAQPage Schema

How do I automate multi-step workflows with required human approval checkpoints?▼

You can automate multi-step workflows with approval checkpoints by running a pipeline that pauses for human validation before executing subsequent actions, ensuring deterministic oversight. This approach is ideal for administrative triage and approval processes.

What is the best way to structure automated pipeline outputs for integration with other systems?▼

The best way to structure automated pipeline outputs for system integration is to enforce a structured JSON enveloped output. This ensures consistency across multi-step workflows and allows downstream systems to easily parse the results.

How do I run an email triage pipeline that requires approval before taking action?▼

To run an email triage pipeline requiring approval, invoke the pipeline command to fetch and classify recent emails, which then pauses execution to require human validation before any automated action is taken.

Does this workflow automation tool support deterministic execution for administrative processes?▼

Yes, this workflow automation tool supports deterministic execution for administrative processes. It enforces structured JSON output and integrates human approval checkpoints to guarantee that repeatable automation actions are explicitly validated.

Can I use this pipeline automation for tasks beyond email triage?▼

Yes, you can use this pipeline automation for any multi-step workflows requiring human oversight. It is ideal for administrative, triage, and approval processes where actions need human validation before execution.

Why does my automated workflow require human validation before executing actions?▼

Your automated workflow requires human validation before executing actions because it is designed with approval checkpoints. This mechanism ensures deterministic execution and prevents errors by requiring human oversight during repeatable automation processes.