amalik/bmad-help

Task: BMAD Help

ROUTING RULES

• Empty phase = anytime — Universal tools work regardless of workflow state
• Numbered phases indicate sequence — Phases like 1-discover → 2-define → 3-build → 4-ship flow in order (naming varies by module)
• Phase with no Required Steps - If an entire phase has no required, true items, the entire phase is optional. If it is sequentially before another phase, it can be recommended, but always be clear with the use what the true next required item is.
• Stay in module — Guide through the active module's workflow based on phase+sequence ordering
• Descriptions contain routing — Read for alternate paths (e.g., "back to previous if fixes needed")
• required=true blocks progress — Required workflows must complete before proceeding to later phases
• Artifacts reveal completion — Search resolved output paths for outputs patterns, fuzzy-match found files to workflow rows

DISPLAY RULES

Command-Based Workflows

When command field has a value:

• Show the command as a skill name in backticks (e.g., bmad-bmm-create-prd)

Skill-Referenced Workflows

When workflow-file starts with skill::

• The value is a skill reference (e.g., skill:bmad-quick-dev), NOT a file path
• Do NOT attempt to resolve or load it as a file path
• Display using the command column value as a skill name in backticks (same as command-based workflows)

Agent-Based Workflows

When command field is empty:

• User loads agent first by invoking the agent skill (e.g., bmad-pm)
• Then invokes by referencing the code field or describing the name field
• Do NOT show a slash command — show the code value and agent load instruction instead

Example presentation for empty command:

Explain Concept (EC)
Load: tech-writer agent skill, then ask to "EC about [topic]"
Agent: Tech Writer
Description: Create clear technical explanations with examples...

MODULE DETECTION

• Empty module column → universal tools (work across all modules)
• Named module → module-specific workflows

Detect the active module from conversation context, recent workflows, or user query keywords. If ambiguous, ask the user.

INPUT ANALYSIS

Determine what was just completed:

• Explicit completion stated by user
• Workflow completed in current conversation
• Artifacts found matching outputs patterns
• If index.md exists, read it for additional context
• If still unclear, ask: "What workflow did you most recently complete?"

EXECUTION

1. Load catalog — Load {project-root}/_bmad/_config/bmad-help.csv

2. Resolve output locations and config — Scan each folder under {project-root}/_bmad/ (except _config) for config.yaml. For each workflow row, resolve its output-location variables against that module's config so artifact paths can be searched. Also extract communication_language and project_knowledge from each scanned module's config.

3. Ground in project knowledge — If project_knowledge resolves to an existing path, read available documentation files (architecture docs, project overview, tech stack references) for grounding context. Use discovered project facts when composing any project-specific output. Never fabricate project-specific details — if documentation is unavailable, state so.

4. Detect active module — Use MODULE DETECTION above

5. Analyze input — Task may provide a workflow name/code, conversational phrase, or nothing. Infer what was just completed using INPUT ANALYSIS above.

6. Present recommendations — Show next steps based on:

   • Completed workflows detected
   • Phase/sequence ordering (ROUTING RULES)
   • Artifact presence

   Optional items first — List optional workflows until a required step is reached Required items next — List the next required workflow

   For each item, apply DISPLAY RULES above and include:

   • Workflow name
   • Command OR Code + Agent load instruction (per DISPLAY RULES)
   • Agent title and display name from the CSV (e.g., "🎨 Alex (Designer)")
   • Brief description

7. Additional guidance to convey:

   • Present all output in {communication_language}
   • Run each workflow in a fresh context window
   • For validation workflows: recommend using a different high-quality LLM if available
   • For conversational requests: match the user's tone while presenting clearly

8. Return to the calling process after presenting recommendations.