Midtrans-Middleware/.bmad/bmm/workflows/1-analysis/research/instructions-router.md

Research Workflow Router Instructions

The workflow execution engine is governed by: {project_root}/.bmad/core/tasks/workflow.xml You MUST have already loaded and processed: {installed_path}/workflow.yaml Communicate in {communication_language}, generate documents in {document_output_language} Web research is ENABLED - always use current {{current_year}} data

🚨 ANTI-HALLUCINATION PROTOCOL - MANDATORY 🚨 NEVER present information without a verified source - if you cannot find a source, say "I could not find reliable data on this" ALWAYS cite sources with URLs when presenting data, statistics, or factual claims REQUIRE at least 2 independent sources for critical claims (market size, growth rates, competitive data) When sources conflict, PRESENT BOTH views and note the discrepancy - do NOT pick one arbitrarily Flag any data you are uncertain about with confidence levels: [High Confidence], [Medium Confidence], [Low Confidence - verify] Distinguish clearly between: FACTS (from sources), ANALYSIS (your interpretation), and SPECULATION (educated guesses) When using WebSearch results, ALWAYS extract and include the source URL for every claim ⚠️ CHECKPOINT PROTOCOL: After EVERY tag, you MUST follow workflow.xml substep 2c: SAVE content to file immediately → SHOW checkpoint separator (━━━━━━━━━━━━━━━━━━━━━━━) → DISPLAY generated content → PRESENT options [a]Advanced Elicitation/[c]Continue/[p]Party-Mode/[y]YOLO → WAIT for user response. Never batch saves or skip checkpoints.

This is a ROUTER that directs to specialized research instruction sets

Check if {output_folder}/bmm-workflow-status.yaml exists No workflow status file found. Research is optional - you can continue without status tracking. Set standalone_mode = true Load the FULL file: {output_folder}/bmm-workflow-status.yaml Parse workflow_status section Check status of "research" workflow Get project_level from YAML metadata Find first non-completed workflow (next expected workflow) Pass status context to loaded instruction set for final update ⚠️ Research already completed: {{research status}} Re-running will create a new research report. Continue? (y/n) Exiting. Use workflow-status to see your next step. Exit workflow ⚠️ Next expected workflow: {{next_workflow}}. Research is out of sequence. Note: Research can provide valuable insights at any project stage. Continue with Research anyway? (y/n) Exiting. Run {{next_workflow}} instead. Exit workflow

Set standalone_mode = false

Welcome {user_name} warmly. Position yourself as their research partner who uses live {{current_year}} web data. Ask what they're looking to understand or research.

Listen and collaboratively identify the research type based on what they describe:

• Market/Business questions → Market Research
• Competitor questions → Competitive Intelligence
• Customer questions → User Research
• Technology questions → Technical Research
• Industry questions → Domain Research
• Creating research prompts for AI platforms → Deep Research Prompt Generator

Confirm your understanding of what type would be most helpful and what it will produce.

Capture {{research_type}} and {{research_mode}}

research_type_discovery

Based on user selection, load the appropriate instruction set

Set research_mode = "market" LOAD: {installed_path}/instructions-market.md Continue with market research workflow Set research_mode = "deep-prompt" LOAD: {installed_path}/instructions-deep-prompt.md Continue with deep research prompt generation Set research_mode = "technical" LOAD: {installed_path}/instructions-technical.md Continue with technical research workflow Set research_mode = "competitive" This will use market research workflow with competitive focus LOAD: {installed_path}/instructions-market.md Pass mode="competitive" to focus on competitive intelligence Set research_mode = "user" This will use market research workflow with user research focus LOAD: {installed_path}/instructions-market.md Pass mode="user" to focus on customer insights Set research_mode = "domain" This will use market research workflow with domain focus LOAD: {installed_path}/instructions-market.md Pass mode="domain" to focus on industry/domain analysis

The loaded instruction set will continue from here with full context of the {research_type}