Skip to content
UXClaim
Product Management

Research-Driven Development

Enforce research-validated development across the full product lifecycle with mandatory validation gates at every phase.

View source More in Product Management

What It Does

Research-Driven Development adds mandatory research and validation gates to every phase of product development. It prevents jumping from ideas directly to implementation by enforcing a structured pipeline: Research → Validate → Plan → Implement → Review → Verify → Deliver.

How It Works

Each phase produces validated artifacts before moving to the next. The skill orchestrates existing development practices while filling critical gaps:

  • Mandatory research before planning or implementation begins
  • Multi-source validation against community best practices and established patterns
  • Research-informed specifications that reference validated findings, not assumptions
  • Quality gates between every phase to prevent advancing with gaps
  • Full lifecycle coverage from discovery through delivery and infrastructure

Use Cases

  • Preventing costly rework by validating technology choices early
  • Ensuring architectural decisions align with existing codebases
  • Building specifications that account for community patterns and best practices
  • Creating accountability for research and decision-making in development workflows
  • Reducing risk in complex product decisions through structured validation

Who Benefits

Product managers coordinating development cycles, engineering leads overseeing quality, and teams implementing AI-assisted coding workflows benefit most. It’s especially valuable when team members make decisions in isolation or when technical choices later prove misaligned with project needs.

Frequently asked questions

How does research-driven development differ from standard development?
Standard workflows often move from idea to implementation quickly. Research-Driven Development inserts mandatory research and validation phases before planning, ensuring decisions are informed by best practices and validated against multiple sources. This prevents costly rework from poor technology choices or architectural misalignment.
What does the validation process actually check?
Each phase's artifacts are validated against community best practices, industry patterns, and existing codebase alignment. Validation confirms research findings are current, technology recommendations fit project constraints, and specifications reference validated findings rather than assumptions.
Can I skip phases or use this incrementally?
No—the skill enforces the full pipeline. Phases cannot be skipped because each builds on the previous phase's validated outputs. However, you can apply it incrementally to new features or modules rather than retrofitting entire projects at once.
How long does the research phase take?
Duration depends on decision complexity. The skill uses ACR complexity scoring to route tasks appropriately and determine research depth. Simple decisions might need minimal research; architectural choices require thorough multi-source validation.
What companion tools work best with this skill?
The skill pairs well with superpowers (for TDD, debugging, and code review) and estimating-agent-tasks (for ACR complexity scoring). It's compatible with Claude Code, Cursor, and other tools supporting the SKILL.md standard format.
How does this integrate with existing development workflows?
It orchestrates existing practices (code review, testing, verification) while adding missing research and validation layers. Your current tools and frameworks continue working; this skill wraps them in a structured lifecycle with mandatory validation gates.
What happens if research contradicts existing assumptions?
The skill surfaces these contradictions explicitly. You can then decide whether to adjust plans, gather additional research, or document the decision rationale. Contradictions are treated as valuable information, not roadblocks.
Is this suitable for fast-moving startup environments?
Yes. While it adds structure, the research phase is scoped by task complexity. Quick decisions move fast; complex architectural choices get thorough research. This reduces later rework costs more than the upfront research investment takes.

Glossary

Quality Gates
Explicit checkpoints between development phases where artifacts must meet validation criteria before proceeding. Gates prevent advancing with gaps or unvalidated decisions.
Multi-Source Validation
Confirming decisions against multiple industry sources, community frameworks, and best practice repositories rather than relying on single references or assumptions.
Research-Informed Specification
A detailed plan or specification document that explicitly references validated research findings, community patterns, and industry best practices rather than making unsupported assumptions.
ACR Complexity Scoring
A system for rating task complexity to determine appropriate research depth and model selection. Simple tasks get lighter research; complex decisions trigger thorough validation.

More in Product Management

All →
Product Management

Agent-Native Design

Design CLIs that work reliably for humans, AI agents, and orchestration systems simultaneously with schema-driven interfaces and safety tiers.

Agents365-ai