FDE Issue Tree Analysis

Use this skill when the user needs to structure and diagnose a complex enterprise delivery problem before jumping into implementation.

When To Use

• The customer says an AI system, knowledge base, agent, automation, or PoC "does not work well" but the root cause is unclear.
• The user needs an Issue Tree, MECE check, root-cause hypotheses, evidence plan, or validation experiments.
• The team is preparing for PoC review, Beta blocker diagnosis, production incident review, or delivery retrospective.
• The problem spans business goals, data, technical chain, process embedding, permissions, security, adoption, or ownership.

How To Use

1. Read README.md for the ISSUE-FDE framework and delivery gates.
2. Use prompt.md for general diagnosis, RAG diagnosis, interview evidence gathering, or MECE self-check.
3. Use checklist.md to verify the tree, hypotheses, evidence plan, and action loop.
4. Use evaluation.md to assess whether the diagnosis is review-ready.
5. Use examples in examples/ when the problem resembles an AI knowledge-base diagnosis.

Output Expectations

Produce a structured diagnosis that includes:

• Problem statement, scope boundary, impact object, and current-stage risk.
• Level 1 to Level 3 Issue Tree with MECE checks.
• Top hypotheses prioritized by impact, evidence availability, and validation cost.
• Evidence requirements and validation experiments.
• FDE action plan with owner, dependency, deliverable, timeline, and regression metric.
• Stage-gate conclusion for PoC, Beta, production, or retrospective.

Do not treat symptoms as root causes. Every leaf node should be testable with data, samples, logs, interviews, or metrics.