Agent skill
nf
Conduct in-depth feature discovery interview to explore, challenge, and document a new feature. Use when asked to 'detail a feature', 'explore a new feature', 'feature discovery', 'interview about feature', or 'spec out a feature'. NOT for quick brainstorming (use /brainstorm), NOT for PRD/JTBD docs (use /product), NOT for implementation tasks (use /ct).
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/nf
SKILL.md
New Feature Discovery
Objective
Conduct a comprehensive interview to fully understand and document a new feature using collaborative brainstorming, then create a formal specification.
Guidelines
- Invoke design-exploration skill first for discovery and exploration
- Use
AskUserQuestiontool for ALL clarifications - provides interactive options for user to choose from - Never assume behavior: if any behavior is unclear/ambiguous (UX flow, edge cases, error handling, states), you MUST ask the user to define expected behavior (preferred via
AskUserQuestion). - Ask non-obvious and thought-provoking questions
- Actively challenge assumptions; do not be a yes-boy. Grill.
- Offer alternatives, shortcuts, and "go deeper" paths
- Continue until the feature is fully understood
- Document everything in the specification file
Workflow
Mobile-app Skill Gate (MANDATORY when feature touches mobile-app/)
Trigger: If the feature will create or modify any code under mobile-app/ (new screens/components, animations, lists, navigation, Expo APIs).
MANDATORY steps (do BEFORE writing recommendations/implementation details in the discovery/spec):
- Read the skill docs:
.claude/skills/react-native-expo-mobile/SKILL.md.claude/skills/react-native-expo-mobile/AGENTS.md
- Explicitly state in the discovery/spec that the skill was read and will be applied in implementation.
- Extract the relevant rules for this feature area (minimum: Core Rendering; plus Animation/List Performance/React Compiler/UI as applicable) and record them under a short section like: "Skill Compliance: react-native-expo-mobile".
Step 1: Codebase Exploration
Launch 1-3 parallel Explore agents (with Sonnet) based on feature complexity. Use Task tool in single message.
Agents:
- Code & patterns in
backend/ - Data models, schemas, APIs (if 2+ agents)
- Docs, ADRs, tests (if 3 agents)
Synthesize findings, then proceed to design exploration.
Step 2: External Research (If Needed)
When you need current information, best practices, or technical research:
- Quick lookups: Use Exa MCP tools (
get_code_context_exa,web_search_exa) directly - In-depth research: Spawn
comprehensive-researchersubagent via Task tool for complex topics requiring multiple sources and cross-verification
Topics to research:
- Industry best practices for the feature type
- Competitor implementations and patterns
- API/library capabilities and limitations
- Security considerations and compliance requirements
Step 3: Design Exploration Phase
Invoke the design-exploration skill with gathered context:
- Ask questions to refine the idea (batch related questions)
- Propose 2-3 different approaches with trade-offs
- Present design incrementally (200-300 word sections) for validation
Step 4: "Grill Me" Challenge Round
After initial design exploration, run a structured challenge:
- Identify assumptions and question each one
- Suggest 2-3 alternative approaches (including a "lean/shortcut" path), if applicable and worth it
- Flag potential risks, hidden costs, or complexity traps
- Recommend where to cut scope vs where to deepen investment
Step 5: Deep-Dive Questions
After design exploration, ask additional non-obvious questions if needed:
Technical Implementation:
- Edge cases and failure scenarios
- Performance implications at scale
- Integration points with existing systems
- Security and privacy implications
UI/UX:
- User mental models and expectations
- Error states and recovery flows
- Accessibility considerations
Business & Product:
- Success metrics and how to measure them
- MVP vs future scope boundaries
- Potential abuse scenarios
Tradeoffs:
- Complexity vs flexibility
- Development speed vs technical debt
- Short-term wins vs long-term maintainability
Step 6: Specification Writing
After design exploration and interview completion:
- Read template: Read
docs/product-docs/templates/discovery-template.mdto load the structure - Create task directory:
tasks/task-YYYY-MM-DD-[feature-name]/ - Write specification following the template structure
- Output file:
discovery-[feature-name].md
- Output file:
- Present summary to user for confirmation
If design exploration reveals the feature is not viable: Document the reasons in a brief discovery-[feature-name]-rejected.md with rationale, and stop here.
Step 7: Codex Validation
After discovery document is created, invoke /codex-cli skill for cross-AI validation.
Verification focus: Discovery document review as senior product analyst
- Completeness, consistency, clarity, feasibility, scope definition
- No conflicts with existing codebase architecture and patterns
- No critical gaps or missing considerations
Process feedback: Update discovery if needed, add "Cross-AI Validation: PASSED" when approved.
If validation fails: Present the issues to the user via AskUserQuestion with options: "Revise discovery doc", "Override and proceed", "Abandon feature".
Output
tasks/task-YYYY-MM-DD-[feature-name]/discovery-[feature-name].md
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?