AgentSkillsCN

issue-solver

用于分析 GitHub 问题、生成解决方案方案并实施所选方法的工作流。适用于处理 GitHub 问题、制定实施计划、对比各类解决方案选项、编辑计划,或落实解决方案的场景。触发条件包括:“解决问题”“修复问题”“实施计划”“编辑计划”“规划格式”“对比选项”。

SKILL.md
--- frontmatter
name: issue-solver
description: Workflow for analyzing GitHub issues, generating solution plans, and implementing chosen approaches. Use when working with GitHub issues, creating implementation plans, comparing solution options, editing plans, or implementing solutions. Triggers on: "solve issue", "fix issue", "implement plan", "edit plan", "plan format", "compare options".

issue-solver Workflow

A two-phase approach to solving GitHub issues with AI assistance.

Quick Reference

TaskCommandAgent
Generate plans (single)/issue-solver:solve 123issue-solver
Generate plans (batch)/issue-solver:solve 1 2 3issue-solver (via Task)
Execute plan/issue-solver:implement 123 2plan-implementer

Parallelism Approaches

ApproachHowBest For
Single issueDirect agent call or /solve NNormal usage
Batch mode/solve 1 2 3 (SDK parallelism)Many similar issues
Outer ClaudeSpawn multiple agentsProgressive results, fault isolation

See multi-issue.md for detailed guidance.

Workflow Phases

Phase 1: Solve (Analysis)

Analyzes an issue and generates 2-4 distinct solution approaches.

Input: GitHub issue number(s) Output: Plan files in plans/issue-{N}/option-{n}-{slug}.md

Parallel Solving: When multiple issues are provided, they are solved concurrently using the Task tool.

Each plan includes:

  • Issue summary
  • Implementation approach
  • Step-by-step instructions
  • Pros/cons and effort estimate

Phase 2: Implement (Execution)

Executes a chosen plan by creating a feature branch and commits.

Input: Issue + option number, OR path to plan file Output: Git commits on a new feature branch

Prerequisites:

  • Clean git working directory
  • Plan file must exist

Plan Files

Plans use YAML frontmatter + markdown. See plan-format.md for specification.

To modify a plan before implementation, see plan-editing.md.

Context Tips

When asking to solve issues, include relevant context:

  • "We use the repository pattern for data access"
  • "Avoid modifying the legacy auth module"
  • "Must maintain Python 3.9 compatibility"

This context helps generate more appropriate solutions.