From external-gitcode-ascend-skills
Analyzes closed GitHub issues to generate structured case studies with root cause analysis and lessons learned. Useful for postmortems and building a knowledge base from issue history.
How this skill is triggered — by the user, by Claude, or both
Slash command
/external-gitcode-ascend-skills:github-issue-summaryThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Analyze closed GitHub issues and generate structured troubleshooting case studies documenting the problem-solving process, root cause, and lessons learned.
Analyze closed GitHub issues and generate structured troubleshooting case studies documenting the problem-solving process, root cause, and lessons learned.
docs/issue-cases/ (configurable)https://github.com/owner/repo/issues/123
or
owner/repo#123
# Using GitHub CLI (recommended)
gh issue view 123 --repo owner/repo --json number,title,body,state,comments,labels,createdAt,closedAt,author
# Get all comments
gh api repos/owner/repo/issues/123/comments
# Get issue events (timeline)
gh api repos/owner/repo/issues/123/timeline
If gh CLI not available, use GitHub API directly:
curl -H "Authorization: token $GITHUB_TOKEN" \
https://api.github.com/repos/owner/repo/issues/123
Generate case studies following this structure:
# Issue #123: [Issue Title]
## Meta
| Field | Value |
|-------|-------|
| Repository | owner/repo |
| Status | Closed |
| Created | YYYY-MM-DD |
| Resolved | YYYY-MM-DD |
| Time to Resolve | X days |
| Labels | bug, priority-high |
## Problem Description
[Original issue body - describe the problem as reported]
## Root Cause Analysis
### Symptoms
- [Observable behavior that indicated the problem]
### Investigation Steps
1. [First debugging step taken]
2. [Second step...]
3. [Key discovery]
### Root Cause
[The actual underlying cause - be specific and technical]
## Resolution
### Fix Applied
[Description of the solution]
### Code/Configuration Changes
```[language]
// Relevant code snippet or config change
[How the fix was verified to work]
## Analysis Guidelines
### Identifying Key Information
1. **Problem Statement**
- First comment (issue body)
- Error messages, stack traces
- Environment details (version, OS, config)
2. **Investigation Trail**
- Follow-up comments with debugging attempts
- Code snippets tested
- Hypotheses raised and tested
- Dead ends explored
3. **Resolution**
- Final solution comment
- Linked PR or commit
- Confirmation of fix
### Key Comments to Extract
Look for comments containing:
- "root cause" / "caused by" / "because"
- "fixed" / "resolved" / "works now"
- Code blocks with solutions
- "try" / "test" / "check" (debugging attempts)
- Error outputs or logs
### Time Analysis
Calculate:
- Time from report to first response
- Time from report to resolution
- Number of back-and-forth exchanges
## Output Location
Default: `docs/issue-cases/issue-{number}-{slug}.md`
Where `slug` is a URL-safe version of the issue title:
"Fix memory leak in worker" → "fix-memory-leak-in-worker"
## Example Usage Scenarios
### Scenario 1: Summarize a specific issue
**User says:** "帮我总结这个issue的定位过程: https://github.com/xxx/yyy/issues/42"
**Actions:**
1. Fetch issue #42 data
2. Analyze comments for troubleshooting steps
3. Extract root cause and resolution
4. Generate case study markdown
5. Save to `docs/issue-cases/issue-42-{title-slug}.md`
### Scenario 2: Batch process recent issues
**User says:** "总结最近10个已关闭的bug类型issue"
**Actions:**
1. Query closed issues with `bug` label
2. For each issue, generate case study
3. Create index file linking all case studies
### Scenario 3: Extract common patterns
**User says:** "分析这些issue的共性问题和解决模式"
**Actions:**
1. Analyze multiple case studies
2. Identify recurring root causes
3. Summarize common debugging patterns
4. Generate pattern summary document
## GitHub CLI Reference
```bash
# View issue details
gh issue view <number> --repo <owner/repo>
# List closed issues
gh issue list --repo <owner/repo> --state closed --limit 20
# Get issue as JSON
gh issue view <number> --repo <owner/repo> --json number,title,body,state,comments,labels
# Get issue timeline
gh api repos/<owner>/<repo>/issues/<number>/timeline
# Search issues
gh issue list --repo <owner/repo> --search "error" --state closed
| Endpoint | Purpose |
|---|---|
GET /repos/{owner}/{repo}/issues/{number} | Issue details |
GET /repos/{owner}/{repo}/issues/{number}/comments | All comments |
GET /repos/{owner}/{repo}/issues/{number}/timeline | Events timeline |
GET /repos/{owner}/{repo}/issues?state=closed | List closed issues |
scripts/fetch_issue.py — Script to fetch issue data via GitHub APIreferences/case-template.md — Full case study templatereferences/analysis-patterns.md — Common troubleshooting patternsnpx claudepluginhub ascend-ai-coding/awesome-ascend-skills --plugin migration-ascend-torchnpu-skillsPerforms root cause analysis on GitHub issues by searching similar issues, analyzing code and comments, and producing a report with probability estimates.
Guides on working with GitHub issues: read descriptions/comments/related PRs, define acceptance criteria with verifiable behaviors, and commit with proper references.
Manages full GitHub issue lifecycle: create with conventional commit titles, sub-issues, cross-repo links, edit/view/list, dump trees to markdown/YAML, push from files, comment/label/close.