Install
$ agentstack add skill-hknc-claude-evolve-root-cause-analysis ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Root Cause Analysis Skill
You systematically diagnose problems to find the underlying cause, not just symptoms.
Philosophy
- Depth over speed - Surface fixes create recurring problems
- Evidence-based - Verify each "why" before going deeper
- Systemic thinking - Look for process/system failures, not just individual errors
- Actionable - End with fixes at the appropriate level
Process
1. Define the Problem Clearly
Start by stating the problem precisely:
- What is happening?
- What should be happening?
- When did it start / how often?
- What's the impact?
**Problem:** API response times increased from 200ms to 3s
**Expected:** Responses under 500ms
**When:** Started Monday after deployment
**Impact:** 40% of users abandoning checkout
2. Choose Analysis Method
| Method | Best For | |--------|----------| | 5 Whys | Linear cause chains, single root cause | | Fishbone | Multiple contributing factors | | Fault Tree | System failures with AND/OR logic | | Timeline | Incidents with sequence of events |
3. Gather Evidence
Before each "why", gather evidence:
For code/systems:
# Logs, metrics, traces
# Use Task tool for parallel investigation if multiple areas
For processes:
- Interview stakeholders
- Review documentation
- Check recent changes
4. Apply 5 Whys
Ask "why" iteratively, but verify each answer:
## 5 Whys Analysis
**Problem:** Users report slow page loads
1. **Why slow?** -> API takes 3 seconds
*Evidence: Network tab shows /api/users taking 3.2s*
2. **Why does API take 3s?** -> Database query is slow
*Evidence: Query logs show SELECT taking 2.8s*
3. **Why is query slow?** -> Full table scan on users table
*Evidence: EXPLAIN shows no index used*
4. **Why no index?** -> Index was dropped in migration
*Evidence: Migration 042 drops idx_users_email*
5. **Why was it dropped?** -> Copied from stack overflow without understanding
*Evidence: Commit message references SO link*
**Root Cause:** Migration review process doesn't catch index changes
5. Identify Fix Levels
Every root cause analysis should identify fixes at multiple levels:
## Fixes
**Immediate (address symptom):**
- Add index back to users table
**Preventive (address root cause):**
- Add migration review checklist
- Add performance regression tests
**Systemic (address deeper issue):**
- Training on database optimization
- Automated index usage analysis in CI
6. Verify Root Cause
Before concluding, verify:
- Does fixing this cause prevent recurrence?
- Are there other contributing factors?
- Is this the deepest actionable cause?
Fishbone Diagram (for multiple factors)
When problem has multiple contributing causes:
┌─ Process: No code review
├─ Process: No testing requirement
People ─────┼─ Skills: Junior dev unfamiliar with DB
└─ Capacity: Team overloaded
Slow ───────────────┼─ Technology: No query monitoring
Page ├─ Technology: No performance budget
Loads Technology ─┼─ Technology: Outdated ORM version
└─
┌─ Environment: Prod DB larger than staging
Environment ┼─ Environment: No realistic test data
└─ Environment: Different DB version
Using Tasks for Parallel Investigation
For complex problems, investigate areas in parallel:
Spawn parallel tasks:
- Task 1: Analyze recent code changes
- Task 2: Check infrastructure metrics
- Task 3: Review error logs
- Task 4: Check external dependencies
Synthesize findings to identify root cause.
Output Format
## Root Cause Analysis: [Problem]
### Problem Statement
[Clear description with impact]
### Analysis Method
[5 Whys / Fishbone / Timeline]
### Investigation
[Step-by-step with evidence]
### Root Cause
[The underlying cause, not the symptom]
### Recommended Fixes
| Level | Action | Effort |
|-------|--------|--------|
| Immediate | [Quick fix] | Low |
| Preventive | [Stop recurrence] | Medium |
| Systemic | [Address deeper issue] | High |
### Verification
[How to confirm the root cause is correct]
Anti-Patterns
DON'T:
- Stop at the first "why" (symptom-level fix)
- Assume without evidence
- Blame individuals (look for system failures)
- Skip verification
DO:
- Keep asking why until you reach an actionable systemic cause
- Verify each step with evidence
- Consider multiple contributing factors
- Propose fixes at multiple levels
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: hknc
- Source: hknc/claude-evolve
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.