القائمة
Hypothesis-driven market demand research with Bayesian probability scoring and strategic analysis. Use for product validation with professional research methodology.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | claw-research |
| description | Hypothesis-driven market demand research with Bayesian probability scoring and strategic analysis. Use for product validation with professional research methodology. |
Violation = Invalid Output - Must Redo
| # | Rule | Check |
|---|---|---|
| 1 | Read sources from config/sources.json only | No hardcoded paths |
| 2 | Every evidence: full_text + source_url + source_type + published_at | 4 fields required |
| 3 | Output analysis-result.json with valid JSON | Schema compliant |
| 4 | Present full Markdown report to user | No "see file" shortcuts |
| 5 | No truncated evidence text | Complete quotes |
| 6 | No placeholder text | No "TBD" or "..." |
CHARTER → SOURCES → COLLECT → ANALYZE → MD_REPORT → NOTION → ACTION
(opt) [BLOCKER] auto AI core [BLOCKER] from MD track
Core Principle: Markdown report is the "Single Source of Truth" (SSOT). Notion sync MUST read from MD file.
Read workspace/projects/{project}/config/research-charter.json if exists.
Run pipeline to collect data:
python3 scripts/run_pipeline.py \
--project-name "my-research" \
--sources "hacker_news,woshipm,pmcaff" \
--query "product pain points"
Key Options:
| Option | Description | Default |
|---|---|---|
--project-name | Unique project identifier | Required |
--sources | Comma-separated sources | All enabled |
--query | Search query | From charter |
--lookback-hours | Time window | 720 (30 days) |
--max-items-per-source | Items per source | 50 |
--include-keyword | Filter: must contain | None |
--exclude-keyword | Filter: must not contain | None |
Output: workspace/projects/{project}/data/candidate_items.jsonl
Collectors output to workspace/projects/{project}/data/:
raw.jsonl - Raw datacandidate_items.jsonl - Filtered itemsRead from data/:
analysis_input.jsoncandidate_items.jsonlresearch-charter.json (if exists)AI Performs:
Must generate Markdown file first as the single source for Notion sync.
workspace/projects/{project}/data/analysis-result.jsonworkspace/projects/{project}/reports/analysis-report-{YYYY-MM-DD}.mdMarkdown File Path:
workspace/projects/{project}/reports/analysis-report-{date}.md
⚠️ Important: This MD file is the ONLY data source for Notion sync. Must contain complete content.
Read content from Markdown file to sync to Notion, ensuring consistency.
┌─────────────────────────────────────────────────────────┐
│ analysis-report-{date}.md ──────► Notion Page │
│ (Single Source of Truth) (Read from MD file) │
└─────────────────────────────────────────────────────────┘
Sync Process:
reports/analysis-report-{date}.md⚠️ Prohibited:
Notion Sync Status:
## Notion Sync Status
- [x] Synced: https://notion.so/page-id
- [ ] SYNC_SKIPPED: [reason - e.g., "Notion MCP not configured"]
Manage action items:
# List pending actions
python3 scripts/manage_actions.py --project "my-research" --list-pending
# Add new action
python3 scripts/manage_actions.py --project "my-research" \
--add "Interview 5 target users" \
--due "2026-04-15" \
--insight "High-volume sellers show strongest pain"
# Complete action
python3 scripts/manage_actions.py --project "my-research" \
--complete "ACTION-001" \
--note "Completed 5 interviews, 3 confirmed willingness to pay"
analysis-result.json){
"generated_at": "UTC timestamp",
"project_name": "string",
"top_pain_points": [{ "label", "summary", "evidence_count", "example_urls" }],
"candidate_clusters": [{ "name", "summary", "evidence_count", "payment_signal" }],
"payment_signals": [{ "label", "summary" }],
"bayesian_scores": {
"success_probability": {
"posterior": 0.35,
"prior": { "value": 0.12, "reasoning": "string" },
"confidence": "high|medium|low",
"signal_assessments": [{ "dimension", "strength", "reasoning", "evidence_quotes", "contribution" }],
"calculation_summary": "string",
"key_uncertainties": ["string"]
},
"payment_probability": { /* same structure */ }
},
"hypothesis_validation": [{ "hypothesis_id", "status", "supporting_evidence", "missing_evidence" }],
"strategic_analysis": {
"swot": { "strengths", "weaknesses", "opportunities", "threats" },
"competitor_landscape": [{ "name", "gap", "our_advantage" }],
"user_journey_stages": [{ "stage", "pain_level", "evidence_count" }]
},
"actionable_insights": [{ "insight", "evidence_basis", "recommended_action", "priority" }],
"action_items": [{ "id", "action", "owner", "status" }],
"strongest_examples": [{ "title", "url", "reason" }],
"risks_or_gaps": ["string"]
}
# [Project] Market Research Report
**Generated**: YYYY-MM-DD | **Confidence**: high/medium/low
## Executive Summary
**Conclusion**: [One sentence]
**Actions**: 1. [P0] ... 2. [P1] ... 3. [P2] ...
**Risks**: [Top 2 risks]
## Bayesian Scores
| Probability | Prior | Posterior | Key Signals |
|-------------|-------|-----------|-------------|
| Success | 12% | 35% | pain_intensity +10% |
| Payment | 6% | 18% | economic_impact +8% |
## Top Pain Points
### Pain 1: [Title] (X evidence)
> **Original**: "[Full quote]"
> **Source**: [Name](URL) | **Type**: community_discussion | **Date**: YYYY-MM-DD
## Strategic Analysis
**SWOT**: Strengths: [evidence] | Weaknesses: [gap] | Opportunities: [trend] | Threats: [risk]
**Competitors**: [Competitor] - Gap: [gap] - Our Advantage: [ours]
## Action Items
- [ ] ACTION-001: [Task] (Due: YYYY-MM-DD)
## Notion Sync
[ ] Synced: [URL] OR [ ] SYNC_SKIPPED: [reason]
Posterior = Prior + Σ(Signal_Strength × Contribution)
| Type | Dimensions |
|---|---|
| Success | pain_intensity, market_evidence, solution_gap, timing_signal, execution_fit |
| Payment | economic_impact, purchase_intent, budget_access, urgency, trust_signals |
Every evidence quote MUST include:
> **Original**: "[Complete original text]"
> **Source**: [Source Name](URL)
> **Type**: community_discussion
> **Published**: YYYY-MM-DD
Source Types: community_discussion | blog_article | industry_survey | competitor_review | documentation | interview | social_post
analysis-result.json exists with valid JSONreports/analysis-report-{date}.md exists with full Markdown reportworkspace/projects/{project}/
├── config/research-charter.json
├── data/
│ ├── raw.jsonl
│ ├── candidate_items.jsonl
│ ├── analysis_input.json
│ └── analysis-result.json
└── reports/
├── analysis-report-{date}.md # ⭐ Single Source of Truth (SSOT)
├── {date}.md # Auto-generated brief report
└── weekly-{year}-{week}.md # Weekly report
config/
├── sources.json # [BLOCKER] Data sources
└── keys.json # API keys (gitignored)
scripts/
├── run_pipeline.py # Main pipeline script
├── manage_actions.py # Action tracker
└── collectors/ # Data collectors
references/
├── analysis-result-schema.md # Detailed JSON schema
└── notion-page-template.md # Report template
┌──────────────────────────────────┐
│ Stage 5: MD Report Generation │
└───────────────┬──────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ reports/analysis-report-{date}.md │
│ ───────────────────────────────── │
│ • Executive Summary │
│ • Bayesian Scores (tables) │
│ • Pain Points (full evidence quotes) │
│ • Strategic Analysis (SWOT, competitors) │
│ • Action Items │
└───────────────────┬─────────────────────────────┘
│
│ READ (Only data source)
│
▼
┌─────────────────────────────────────────────────┐
│ Stage 6: Notion Sync │
│ ───────────────────── │
│ • Parse MD file structure │
│ • Convert to Notion block format │
│ • Preserve content integrity (no simplification)│
└───────────────────┬─────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ Notion Page │
│ ────────── │
│ Content == MD File Content ✓ │
└─────────────────────────────────────────────────┘
⚠️ Consistency Guarantee: