deep-research（深入研究）

[BLOCKING] Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval. [BLOCKING] Before each step or sub-skill call, update task tracking: set in_progress when step starts, set completed when step ends. [BLOCKING] Every completed/skipped step MUST include brief evidence or explicit skip reason. [BLOCKING] If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions.

[IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting.

External Memory: For complex or lengthy work (research, analysis, scan, review), write intermediate findings and final results to a report file in plans/reports/ — prevents context loss and serves as deliverable.

Evidence Gate: MANDATORY IMPORTANT MUST ATTENTION — every claim, finding, and recommendation requires file:line proof or traced evidence with confidence percentage (>80% to act, <80% must verify first).

Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act. Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.

AI Mistake Prevention — Failure modes to avoid on every task: - Check downstream references before deleting. Deleting components causes documentation and code staleness cascades. Map all referencing files before removal. - Verify AI-generated content against actual code. AI hallucinates APIs, class names, and method signatures. Always grep to confirm existence before documenting or referencing. - Trace full dependency chain after edits. Changing a definition misses downstream variables and consumers derived from it. Always trace the full chain. - Trace ALL code paths when verifying correctness. Confirming code exists is not confirming it executes. Always trace early exits, error branches, and conditional skips — not just happy path. - When debugging, ask "whose responsibility?" before fixing. Trace whether bug is in caller (wrong data) or callee (wrong handling). Fix at responsible layer — never patch symptom site. - Assume existing values are intentional — ask WHY before changing. Before changing any constant, limit, flag, or pattern: read comments, check git blame, examine surrounding code. - Verify ALL affected outputs, not just the first. Changes touching multiple stacks require verifying EVERY output. One green check is not all green checks. - Holistic-first debugging — resist nearest-attention trap. When investigating any failure, list EVERY precondition first (config, env vars, DB names, endpoints, DI registrations, data preconditions), then verify each against evidence before forming any code-layer hypothesis. - Surgical changes — apply the diff test. Bug fix: every changed line must trace directly to the bug. Don't restyle or improve adjacent code. Enhancement task: implement improvements AND announce them explicitly. - Surface ambiguity before coding — don't pick silently. If request has multiple interpretations, present each with effort estimate and ask. Never assume all-records, file-based, or more complex path.

Quick Summary

Goal: Deep-dive into top sources, extract key findings, cross-validate claims, build structured evidence base.

Workflow:

1. Read source map — Load output from web-research step
2. Fetch top sources — WebFetch top 5-8 Tier 1-2 sources
3. Extract findings — Pull key facts, data points, quotes
4. Cross-validate — Compare findings across sources
5. Build evidence base — Structured findings with confidence scores

Key Rules:

• Maximum 8 WebFetch calls per invocation
• Every finding must cite specific source
• Conflicting claims → present both, flag discrepancy

Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).

Deep Research

Step 1: Load Source Map

Read the source map from .claude/tmp/_sources-{slug}.md (output of web-research step).

Prioritize sources for deep-dive:

1. Tier 1-2 sources first
2. High-relevance sources
3. Sources covering identified gaps

Step 2: Fetch Top Sources

For each priority source (max 8):

1. Run WebFetch with the URL
2. Extract: key claims, data points, quotes, methodology
3. Note: publication date, author credentials, source type

Step 3: Extract Findings

For each source, extract:

• Key claims — factual statements with specific data
• Data points — numbers, percentages, dates
• Quotes — notable expert statements
• Methodology — how data was gathered (for market reports)

Step 4: Cross-Validate

Compare findings across sources:

• Agreement — 2+ sources say the same thing → high confidence
• Discrepancy — sources disagree → note both positions
• Unique — only 1 source → mark as "single source, unverified"

Step 5: Build Evidence Base

Write to .claude/tmp/_evidence-{slug}.md:

# Evidence Base: {Topic}

**Date:** {date}
**Sources analyzed:** {count}

## Findings

### Finding 1: {Title}

**Confidence:** {95%|80%|60%|<60%}
**Sources:** [1], [3]
**Content:** {finding with inline citations}
**Cross-validation:** {agreement/discrepancy notes}

## Unresolved Discrepancies

- {claim X from source A vs claim Y from source B}

## Gaps Remaining

- {what couldn't be verified}

Workflow Recommendation

MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS: If you are NOT already in a workflow, you MUST ATTENTION use AskUserQuestion to ask the user. Do NOT judge task complexity or decide this is "simple enough to skip" — the user decides whether to use a workflow, not you: 1. Activate research workflow (Recommended) — web-research → deep-research → synthesis → review 2. Execute /deep-research directly — run this skill standalone

Next Steps

MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS after completing this skill, you MUST ATTENTION use AskUserQuestion to present these options. Do NOT skip because the task seems "simple" or "obvious" — the user decides:

• "/business-evaluation (Recommended)" — Evaluate business viability from research
• "/knowledge-synthesis" — If synthesizing research report
• "Skip, continue manually" — user decides

Closing Reminders

MANDATORY IMPORTANT MUST ATTENTION break work into small todo tasks using TaskCreate BEFORE starting. MANDATORY IMPORTANT MUST ATTENTION validate decisions with user via AskUserQuestion — never auto-decide. MANDATORY IMPORTANT MUST ATTENTION add a final review todo task to verify work quality.

• MUST ATTENTION apply critical thinking — every claim needs traced proof, confidence >80% to act. Anti-hallucination: never present guess as fact.
• MUST ATTENTION apply AI mistake prevention — holistic-first debugging, fix at responsible layer, surface ambiguity before coding, re-read files after compaction.

[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using TaskCreate.

[IMPORTANT] Analyze how big the task is and break it into many small todo tasks systematically before starting — this is very important.

Prompt-Enhance Closing Anchors

• IMPORTANT MUST ATTENTION follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval
• IMPORTANT MUST ATTENTION for every step/sub-skill call: set in_progress before execution, set completed after execution
• IMPORTANT MUST ATTENTION every skipped step MUST include explicit reason; every completed step MUST include concise evidence
• IMPORTANT MUST ATTENTION if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses