diagnostic-action
Use when the user asks why something happened, what’s wrong, root cause analysis, debugging, or incident explanation.
When & Why to Use This Skill
The Diagnostic Action skill is a specialized AI framework designed for systematic root cause analysis and technical troubleshooting. It helps users move beyond symptoms to identify the underlying causes of software bugs, system failures, or unexpected incidents by generating ranked hypotheses based on evidence and logical mechanisms, significantly accelerating the debugging process.
Use Cases
- Software Debugging: Analyzing error logs, stack traces, and unexpected behaviors to pinpoint the exact source of a code failure.
- Root Cause Analysis (RCA): Conducting a structured investigation into system outages or performance regressions to understand 'why' an incident occurred.
- Failure Analysis: Investigating why a specific automated workflow or business process failed to produce the expected results.
- Hypothesis Testing: Systematically evaluating multiple potential causes for intermittent or complex technical issues to narrow down the most likely culprit.
- Incident Explanation: Providing clear, technical explanations for stakeholders regarding the mechanics of a recent system failure or bug.
| name | diagnostic-action |
|---|---|
| description | Use when the user asks why something happened, what’s wrong, root cause analysis, debugging, or incident explanation. |
| version | 1.0 |
Diagnostic Action Skill
Purpose
Identify and explain likely causes of an observed problem or symptom.
When to use
- “Why did this happen?”
- Root cause analysis
- Debugging / incident investigation
- Failure analysis
Do NOT use for:
- recommendations or fixes (use prescriptive-actions)
- plans or procedures (use planning-action / procedural-action)
Operating rules
- Restate symptoms and scope clearly.
- Generate multiple plausible hypotheses.
- Explain the mechanism for each hypothesis.
- Rank by likelihood and impact.
- Identify missing data that would disambiguate causes.
- Do not prescribe fixes unless explicitly asked.
Outputs
Symptoms & scope
Hypotheses (ranked)
- Cause
- Mechanism
- Evidence for/against
- Confidence