How do I install the Trace skill?

Run npx skillmds@latest add majiayu000/trace-8 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Trace safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Trace?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Trace free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Trace by majiayu000 — Agent Skill

Name: Trace
Rating: 4.2 (5 reviews)
Author: majiayu000

File contents

trace

Root cause analysis when surface fixes fail.

When to Use

Trigger	Action
Bug persists after initial fix	Run trace
Production incident (SEV 1-2)	Run trace
Complex failure (multiple sources)	Run trace
Trivial bug (< 10 min fix)	Skip

1. Timeline

HH:MM - [Event] - [System state] - [Action taken]

Note: detection time vs. start time (how long hidden?)

2. Five Whys

Effect: [What users/systems experienced]

Why 1: [Immediate cause] (X-Y% confident)
Why 2: [Underlying mechanism] (X-Y% confident)
Why 3: [System/process gap] (X-Y% confident)
Why 4: [Organizational factor] (X-Y% confident)
Why 5: [Root cause] (X-Y% confident)

Gates: < 70% any level → request logs/reproduction Root cause = deepest answer with ≥70% confidence

3. Contributing Factors

Beyond root cause, what amplified impact?

Process: Monitoring gaps, testing gaps, review gaps People: Unclear ownership, missing runbooks Technical: Dependency failures, capacity limits, config drift Context: Traffic patterns, deployment timing, external factors

List 2-4 factors with specific mechanisms.

Tools: Ishikawa for categorization, Iceberg for deep structure analysis.

4. Impact

Users: Count, experience, business cost (quantified) Systems: Downstream effects, data integrity, security

5. Prevention Hierarchy

Immediate (< 1 week)

1. [Code/config change]
   File: [path:line]
   Verification: [how to confirm]
   Owner: [who]

Short-Term (< 1 month)

1. [Alert/test/automation]
   Trigger: [when it fires]
   Owner: [who]

Long-Term (< 1 quarter)

1. [Architectural change]
   Problem class: [what it prevents]
   Effort: [story points]
   Owner: [who]

Each tier needs ≥1 item.

6. Blameless

See soul/references/blameless.md

Focus on system gaps, not individual mistakes.

Output

# Incident: [Title]

## Timeline
[Detection → Resolution]

## Five Whys
[With confidence per level]

## Contributing Factors
[2-4 items]

## Impact
[Quantified]

## Prevention
[Immediate / Short-Term / Long-Term]

## Top 3 Actions
1. [Task] | Owner: [Name] | Due: [Date]
2. ...
3. ...

Anti-Patterns

Bad	Good
"Human error"	Identify automation gap
"Communication failure"	Identify process gap
"Be more careful"	Add system safeguard
"Performance issue"	Specific: "Pool exhausted at 1K req/s"

Boundary

System gaps only. If this feels personal, stop.

Never assign blame to individuals
If analysis points to a person, reframe to the system that allowed it
User interprets organizational meaning — Claude identifies patterns

Trace finds gaps, not fault.

---name: trace-8description: Root cause analysis for complex bugs. Use when initial fix fails or incident is severe. Traces Effect → Cause → Root with confidence levels and prevention hierarchy.---# traceRoot cause analysis when surface fixes fail.## When to Use| Trigger | Action || ---------------------------------- | --------- || Bug persists after initial fix | Run trace || Production incident (SEV 1-2) | Run trace || Complex failure (multiple sources) | Run trace || Trivial bug (< 10 min fix) | Skip |## 1. Timeline```HH:MM - [Event] - [System state] - [Action taken]```Note: detection time vs. start time (how long hidden?)## 2. Five Whys```Effect: [What users/systems experienced]Why 1: [Immediate cause] (X-Y% confident)Why 2: [Underlying mechanism] (X-Y% confident)Why 3: [System/process gap] (X-Y% confident)Why 4: [Organizational factor] (X-Y% confident)Why 5: [Root cause] (X-Y% confident)```**Gates**: < 70% any level → request logs/reproductionRoot cause = deepest answer with ≥70% confidence## 3. Contributing FactorsBeyond root cause, what amplified impact?**Process**: Monitoring gaps, testing gaps, review gaps**People**: Unclear ownership, missing runbooks**Technical**: Dependency failures, capacity limits, config drift**Context**: Traffic patterns, deployment timing, external factorsList 2-4 factors with specific mechanisms.**Tools:** [Ishikawa](../soul/references/tools/ishikawa.md) for categorization, [Iceberg](../soul/references/tools/iceberg.md) for deep structure analysis.## 4. Impact**Users**: Count, experience, business cost (quantified)**Systems**: Downstream effects, data integrity, security## 5. Prevention Hierarchy### Immediate (< 1 week)```1. [Code/config change] File: [path:line] Verification: [how to confirm] Owner: [who]```### Short-Term (< 1 month)```1. [Alert/test/automation] Trigger: [when it fires] Owner: [who]```### Long-Term (< 1 quarter)```1. [Architectural change] Problem class: [what it prevents] Effort: [story points] Owner: [who]```Each tier needs ≥1 item.## 6. BlamelessSee `soul/references/blameless.md`Focus on system gaps, not individual mistakes.## Output```# Incident: [Title]## Timeline[Detection → Resolution]## Five Whys[With confidence per level]## Contributing Factors[2-4 items]## Impact[Quantified]## Prevention[Immediate / Short-Term / Long-Term]## Top 3 Actions1. [Task] | Owner: [Name] | Due: [Date]2. ...3. ...```## Anti-Patterns| Bad | Good || ----------------------- | -------------------------------------- || "Human error" | Identify automation gap || "Communication failure" | Identify process gap || "Be more careful" | Add system safeguard || "Performance issue" | Specific: "Pool exhausted at 1K req/s" |---## Boundary**System gaps only. If this feels personal, stop.**- Never assign blame to individuals- If analysis points to a person, reframe to the system that allowed it- User interprets organizational meaning — Claude identifies patternsTrace finds gaps, not fault.

Frequently asked questions

Trace