Root Cause Incident Workflow

What this skill does

Guide incidents and repeated failures through root-cause diagnosis before fixes or retries.

It is designed for buyers who need a repeatable, auditable workflow instead of a one-off prompt. The skill asks for the minimum useful inputs, produces a structured deliverable, and includes quality gates so the result can be reviewed or handed to another teammate.

Best buyer input

Problem statement, symptoms, logs/errors, recent changes, affected systems, and attempted fixes.

Output contract

A root-cause report with hypotheses, evidence table, test plan, fix options, retry guardrails, and verification checklist.

Every run should include:

1. A short intake summary.
2. The recommended workflow or deliverable structure.
3. Evidence, assumptions, or missing-input notes.
4. A pass/fail quality checklist.
5. A next-action section the buyer can execute immediately.

Workflow

1. Clarify the buyer goal, target audience, and constraints.
2. Convert the input into a structured working brief.
3. Build the primary deliverable using the required sections from the output contract.
4. Check for gaps, contradictions, overclaims, missing data, and execution risks.
5. Return the final packet in a clean, client-ready format with a concise next step.

Quality bar

• Do not invent metrics, sources, screenshots, credentials, or platform access.
• Mark assumptions clearly when buyer input is incomplete.
• Keep private account data, access credentials, and internal-only implementation details out of the final result.
• Prefer concrete checklists, tables, and action steps over vague advice.
• Stop and ask for the missing input only when it would materially change the deliverable.

Example request

"Use this source material to create a root cause incident workflow packet for a small team that needs a decision-ready deliverable by tomorrow."

Example result shape

• Intake summary
• Recommended structure
• Main deliverable
• QA checklist
• Risks and assumptions
• Next actions