A high-integrity development framework that dispatches fresh subagents per task with mandatory two-stage reviews to ensure spec compliance and code quality.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install lobster-subagent-dev
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install lobster-subagent-dev using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Lobster Subagent-Driven Development?
Lobster Subagent-Driven Development is a specialized methodology within the Openclaw Skills library designed to scale AI-assisted development through isolation and rigorous verification. By spinning up a fresh subagent for every individual task, it prevents context drift and ensures that AI models operate on clean, relevant information. This approach treats subagents as ephemeral workers, allowing the primary agent to maintain a high-level coordination role without inheriting messy session histories.
The skill implements a robust review pipeline inspired by Jesse Vincent's Superpowers, ensuring that every line of code is checked against the original specification and general quality standards before being merged. This structure is ideal for complex engineering projects where accuracy and maintainability are non-negotiable.
Lobster Subagent-Driven Development Use Cases
- Implementing complex features from a multi-task technical plan
- Maintaining high code standards in large-scale Openclaw Skills projects
- Reducing AI hallucinations by resetting subagent context for every task
- Scaling development via structured, multi-stage AI reviews
- Cost-optimizing development by routing mechanical tasks to cheaper models
How Lobster Subagent-Driven Development Works
- Plan Extraction: The system reads the implementation plan and decomposes it into a comprehensive list of discrete tasks tracked via TodoWrite.
- Implementer Dispatch: A subagent is assigned a specific task, provided with necessary context, and tasked to implement, test, and commit the changes.
- Spec Review: A second subagent verifies that the implementation exactly matches the provided functional requirements.
- Quality Review: A third subagent evaluates the code for performance, readability, and security standards.
- Task Finalization: Once both reviews pass, the task is marked complete, and the workflow moves to the next item in the queue.
Lobster Subagent-Driven Development Setup
To integrate this workflow into your Openclaw Skills environment, ensure your agent has subagent dispatch capabilities enabled.
# Install the skill to your local registry
openclaw install lobster-subagent-dev
# Configure model routing based on task complexity
openclaw configure lobster-subagent-dev --model-routing=dynamic
Lobster Subagent-Driven Development Data Schema & Taxonomy
The skill manages task states and agent transitions using a structured status table:
| Status | Definition | Next Action |
|---|---|---|
| DONE | Task completed without issues | Proceed to spec review |
| DONE_WITH_CONCERNS | Task completed but has caveats | Manual intervention or re-review |
| NEEDS_CONTEXT | Subagent requires more data | Provide info and re-dispatch |
| BLOCKED | Task cannot proceed | Escalation or model swap |
All progress is tracked through a persistent TodoWrite log to ensure state recovery across sessions.
Lobster Subagent-Driven Development Advanced Features
- Cost-Aware Routing: Automatically selects the most economical model for mechanical tasks while reserving flagship models for architecture reviews.
- Strict Multi-Stage Verification: Enforces a no-skip policy on spec and quality reviews to prevent technical debt.
- Context Isolation: Ensures subagents never inherit messy session context, maximizing the reliability of their output.
- Conflict Prevention: Native logic prevents parallel implementer dispatch to avoid merge conflicts and file corruption.
- Final Global Review: Triggers a holistic codebase audit after all individual tasks are completed to ensure integration integrity.
SKILL.md
Loading
Related Openclaw Skills
YouTube Downloader & Skimmer
A specialized tool to download YouTube videos and automatically slice them into segments based on chapters or custom timestamps.
Lobster Debugging
A rigorous 4-phase debugging framework for AI agents that mandates systematic root-cause analysis over temporary symptom patching.
Search First
A systematic research-before-coding workflow that prevents reinventing the wheel by prioritizing existing tools, libraries, and MCP servers.
Feishu Cron Delivery for OpenClaw
A specialized skill for configuring and repairing reliable scheduled message delivery from OpenClaw agents to Feishu.
Anthropic Cost Optimizer
A specialized audit tool that rewrites your OpenClaw configuration to drastically reduce Anthropic API expenses through intelligent model routing and prompt caching.
AI Stock Investment Advisor
A comprehensive AI-driven advisory system that automates global stock data collection, multimodal technical analysis, and professional investment report generation.
Featured*
