Menu
Deep research and architecture design for new product development
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Structured task execution with quality gates, parallel waves, and pm-db tracking — solo or with multi-agent teams (--team). Use this skill when the user wants to execute a task list, run implementation tasks, start building a feature from a plan, run tasks in parallel with a team of agents, or says things like 'execute these tasks', 'start building', 'run the implementation', 'execute the plan', 'execute with a team', or 'start phase execute'. Also triggers on '/start-phase execute' and '/start-phase-execute-team' commands.
Assess and score codebase architecture health — layer separation, SOLID compliance, module coupling (FAN-IN/FAN-OUT), circular dependencies, drift from documented patterns. Use whenever the user asks to assess/review/grade architecture quality, wants an architecture score, or asks how healthy the codebase's architecture is. Analysis-only report + refactoring tasks; for docs-vs-code drift use document-hub-analyze.
Deep analysis of codebase for code duplication. Detects exact, structural, and pattern-level duplicates, generates comprehensive reports with refactoring suggestions and metrics.
Use when Mark reports a bug by pasting a raw artifact with little or no prose — a stack trace, an HTTP request/response trace (URL + method + status like 500/403/404/502), a console/React/Prisma error, a ChunkLoadError, a shell error, or a screenshot of a visual defect — or when he says things like 'still failing', 'clicking X does nothing', 'no inline editor on the rendered page', 'do we have local postgres/pgvector?', or 'identify the true error'. This is the front-to-back debugging procedure — extract the trigger and reproduction from terse input, probe whether the underlying plumbing even exists, localize by contrasting known-good vs broken, and drive to a durable root cause rather than a workaround. Reach for this BEFORE proposing any fix.
Deep analysis of codebase vs documentation alignment (cline-docs/). Detects drift, identifies undocumented code, extracts missing glossary terms, and provides actionable recommendations without making changes. Use this skill when the user asks "are the docs up to date", "check documentation quality", "what's missing from the docs", or wants a read-only audit before deciding what to update. For actually making changes, use document-hub-update instead.
Bootstrap a new project's documentation hub by creating the four core documentation files (systemArchitecture.md, keyPairResponsibility.md, glossary.md, techStack.md) with initial content based on codebase analysis.
| name | new-product |
| description | Deep research and architecture design for new product development |
| args | {"input":{"type":"string","description":"Path to documentation files or product description","required":true}} |
Given documentation or a product description, perform deep research and cross-reference technologies to determine the optimal architecture for building a new application.
You are orchestrating a 6-stage deep research workflow to design product architecture. Run the stages in order, use 4 parallel agents for the deep-research stages, and present a final comprehensive review at the end.
All document and agent-prompt templates live in this skill's references/templates/ — read the listed template at each stage; do not improvise document structures.
/ or ends in .md, .txt, etc.), read the documentation; otherwise treat it as a product description.realtime-chat). Store it for later stages.mkdir -p /job-queue/product-{name}/research-notesAskUserQuestion — Product Vision, Scale & Audience, Deployment Context. Exact question/option text: references/templates/stage1-big-idea.md.WebSearch: "[product type] architecture patterns 2026", "[product type] technology stack comparison 2026", "[product type] best practices 2026", "[scale level] architecture considerations 2026". WebFetch the top 2-3 articles per search; extract common patterns, recommended technologies, pitfalls. Save notes to /job-queue/product-{name}/research-notes/stage1-research.md./job-queue/product-{name}/big-idea.md using the content structure in references/templates/stage1-big-idea.md (overview, value proposition, architecture approach, frontend/backend/database/deployment options with pros/cons, key challenges, next steps, research sources).Launch all 4 research agents at once via the Task tool with subagent_type="general-purpose". Each agent researches one architectural document independently.
| Stage | Task description | Output document | Prompt template |
|---|---|---|---|
| 2 | "Runtime execution research" — how the system executes work | runtime-execution.md | references/templates/stage2-runtime-execution.md |
| 3 | "Abstraction layer research" — how user intent becomes executable logic | abstraction-layer.md | references/templates/stage3-abstraction-layer.md |
| 4 | "Integration layer research" — how the system connects to external resources | integration-layer.md | references/templates/stage4-integration-layer.md |
| 5 | "Output rendering research" — how results are delivered to consumers | output-rendering.md | references/templates/stage5-output-rendering.md |
Build each agent prompt from its template:
references/templates/common-closing-sections.md — the single deduplicated tail (Technology Comparison / Implementation Considerations / Research Sources) shared by all four documents; insert it at the template's [COMMON CLOSING SECTIONS] marker with that stage's variants.[Product Name] and {name}, and paste the full big-idea.md content into the CONTEXT block (every agent needs it for cross-referencing).Each template already instructs the agent to WebSearch/WebFetch, ask its own AskUserQuestion clarifications, and write both its document and raw notes under /job-queue/product-{name}/.
Wait for all 4 agents to complete, then display "✅ Stage N/6: [topic] research complete" for stages 2-5.
/job-queue/product-{name}/ARCHITECTURE-SUMMARY.md (stack, style, key decisions, trade-offs, risks, next steps) and print the completion banner — both templates in references/templates/stage6-summary-and-review.md.AskUserQuestion (exact text in the same template): Approve, or Revise runtime execution / abstraction layer / integration layer / output rendering.
On any stage failure: display which stage failed, list completed vs failed stages, show the error, note partial work saved at /job-queue/product-{name}/, and give recovery instructions (resume with /new-product [input] or continue manually). Example failure display: references/templates/stage6-summary-and-review.md.
/job-queue/product-{name}/ (big-idea, 4 architecture docs, ARCHITECTURE-SUMMARY), trade-offs and risks identified, user approval obtained.