A sophisticated fan-out skill that coordinates concurrent AI subagents by clustering independent tasks, drafting complete tickets, and conducting final integrations.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skillet-stations
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 skillet-stations 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 stations?
The stations skill acts as an automated expeditor for concurrent AI agents. When faced with multiple bugs, independent subsystems, or separate repositories, this skill handles the triage and fan-out process. Instead of blindly dispatching an agent for every individual symptom or failing test, it enforces a strict discipline of proving task independence before parallelization.
By leveraging this component of Openclaw Skills, developers and AI orchestrators can prevent conflicting local symptom patches. It ensures that tasks sharing a root cause are grouped into a single station, while truly independent workloads run concurrently in isolated scopes or separate worktrees. This systematic approach eliminates integration bottlenecks and keeps codebases architecturally coherent.
stations Use Cases
- Handling explicit user prompts to parallelize, fan out, or execute multiple tasks simultaneously.
- Triage and resolution of an extensive list of unrelated CI/CD failures or test suite crashes.
- Splitting large development initiatives into concurrent, file-disjoint subsystems or micro-features.
- Managing complex refactoring efforts that span across multiple independent repositories.
How stations Works
- Triage and Clustering: The skill analyzes symptoms and recent git history to cluster tasks by suspected root cause rather than individual files.
- Independence Verification: It checks write sets to guarantee that concurrent stations do not edit overlapping files, enforcing serialization or creating dedicated worktrees where necessary.
- Ticket Generation: A completely isolated, contextual ticket is drafted for each subagent station containing explicit scopes, raw symptoms, and strict output contracts.
- Execution and Monitoring: Concurrent AI subagents are dispatched to execute their isolated workloads without inheriting unnecessary global state.
- Integration and Review: Upon subagent completion, the skill reviews git diff stats, checks for semantic cross-cutting conflicts, and executes a full integrated test suite run.
stations Setup
To activate and configure the stations skill within your environment, include it in your skill configurations and ensure the base environment supports worktree isolation.
# Initialize or check status of openclaw skills in your workspace
openclaw skills list
# Ensure dependent skills like check and refire are accessible
openclaw skills enable stations
No specialized third-party dependencies are required, but a functioning git CLI environment is highly recommended to support write-set verification.
stations Data Schema & Taxonomy
The stations skill utilizes a highly structured tracking schema to manage the lifecycles of concurrent subagents. It generates individual tickets and validates integration outputs based on the following structural model:
| Object | Field | Type | Description |
|---|---|---|---|
| Ticket | scope |
String | Named file, test file, or subsystem allowed to be modified. |
symptom |
String | Raw error logs, failing test names, and expected vs observed behavior. | |
context |
String | Documented contracts, historical diffs, and adjacent working code. | |
constraints |
Array | Boundaries such as prohibiting test rewriting or out-of-scope edits. | |
| Result | diff_stat |
String | Verification of actual modified files using git diff --stat. |
output_contract |
String | Validation logs and cross-cutting architectural observations. |
All temporary file-state discrepancies or manual worktree paths generated during parallel execution are dynamically added to the project's local .gitignore rules.
stations Advanced Features
- Root-Cause Clustering Protection: Automatically merges overlapping symptom reports into a single station to prevent contradictory localized fixes.
- Write-Set Conflict Prevention: Enforces automated serialization or spins up native harness worktrees if concurrent agents attempt to edit overlapping files.
- Semantic Discord Auditing: Analyzes individual green subagent responses for underlying data shape or architectural mismatches prior to main-branch integration.
- Isolation Enforcement: Restricts subagent access to global contexts, making prompts the sole source of truth to reduce token overhead and hallucinatory behavior.
SKILL.md
Loading
Related Openclaw Skills
Todo.txt Task Manager
A command-line task management integration utilizing the todo.txt format to organize projects, priorities, and contexts directly through an AI agent.
Special
The Special skill is a read-only architecture discovery tool that analyzes your repository's existing codebase to propose concrete, high-leverage feature roadmaps and opportunities.
Image Analyzer Advisor
An AI-driven image analysis skill designed to inspect visual elements, composition, and quality to provide actionable recommendations.
MiniMax TTS AI Voice Synthesis
A command-line tool and integration script that synthesizes ultra-realistic, multilingual text-to-speech using the high-performance MiniMax HTTP REST API.
skillify
A utility designed to capture manual workflows, transcripts, and command patterns, converting them into discoverable, highly effective agent skills.
Sendback
The sendback skill provides a systematic protocol for AI coding agents to critically analyze, verify, and respond to human or bot code reviews instead of implementing feedback on reflex.
Featured*
