playbooks

home / skills / anton-abyzov / specweave / github-task-splitter

github-task-splitter skill

/plugins/specweave-github/skills/github-task-splitter

This skill helps split SpecWeave tasks across multiple GitHub repositories by mapping microservices and cross-repo dependencies.

This is most likely a fork of the sw-github-task-splitter skill from openclaw
npx playbooks add skill anton-abyzov/specweave --skill github-task-splitter

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
468 B
---
name: github-task-splitter
description: Expert agent for splitting SpecWeave tasks across multiple GitHub repositories based on microservice or multi-repo architecture. Use when tasks span multiple services, creating cross-repo tracking issues, or analyzing dependencies across repositories.
model: opus
context: fork
---

# GitHub Task Splitter Agent

Expert agent for splitting SpecWeave tasks across multiple GitHub repositories based on architecture patterns.

Overview

This skill is an expert agent that splits SpecWeave tasks across multiple GitHub repositories to fit microservice or multi-repo architectures. It automates creation of cross-repo tracking issues, maps dependencies, and generates per-repo work items so teams can act independently while preserving end-to-end traceability. The agent is implemented in TypeScript and integrates with common developer tools and workflows.

How this skill works

The agent analyzes a SpecWeave task and the target repository graph to identify logical service boundaries, shared libraries, and dependent components. It then generates repo-specific subtasks, creates cross-repo tracking issues on GitHub, and annotates dependencies so each repository has clear acceptance criteria and links back to the parent task. Optional integrations let it surface tasks to Azure DevOps, Jira, or testing pipelines.

When to use it

  • A task touches multiple microservices or repositories
  • You need to create cross-repo tracking and dependency links
  • Coordinating parallel work across teams with independent repositories
  • Breaking up a large spec into implementable per-repo tickets
  • Preparing multi-repo test and deployment plans

Best practices

  • Provide a clear SpecWeave task with service ownership details to improve split accuracy
  • Define repository boundaries and shared library owners in advance
  • Review generated per-repo issues for acceptance criteria and priorities
  • Add explicit dependency metadata so CI/CD and reviewers can sequence work
  • Use the agent iteratively: refine splits after an initial pass to catch missed edges

Example use cases

  • Split a feature request that alters API contracts across an auth service, payments service, and gateway into per-repo tasks with a coordinating tracking issue
  • Create issues for migrating a shared utility library used by multiple repos and list migration order and fallout
  • Analyze a cross-cutting change (e.g., observability or config) and produce implementation tasks and test scopes per repository
  • Prepare multi-repo release tasks with linked testing and rollback steps for each service
  • Generate Jira or Azure DevOps work items from the per-repo splits and include links back to the original SpecWeave spec

FAQ

Can the agent update CI/CD or only create issues?

It can create issues and annotate dependency metadata; with configured integrations it can also update CI/CD pipelines and create corresponding work items in Jira or Azure DevOps.

How does it detect repository boundaries?

It uses repository metadata, dependency graphs, and hints from the SpecWeave task (service names, ownership) to infer logical boundaries, and supports manual overrides.