feature-analyzer

Feature Analyzer

Load and analyze all feature artifacts to build comprehensive understanding for downstream operations.

Workflow Overview

1. Scan - Check artifact availability
2. Load - Read all feature documents
3. Build - Construct complete mental model
4. Confirm - Ready for operations

Step 1: Scan Feature Directory

Run prerequisites scanner to identify available artifacts:

.claude/​skills/​feature-analyzer/​scripts/​check-prerequisites.sh <feature-directory>

Scanner checks for:

• Core artifacts: spec.md, ux.md, plan.md, tasks.md
• Data artifacts: data-model.md, contracts/
• Support artifacts: research.md, setup.md

Outputs JSON with AVAILABLE and MISSING arrays. Exits with error if core files missing.

Step 2: Load Feature Artifacts

Read all available artifacts in priority order:

Core Documents (Required)

• spec.md - Requirements (FR-XXX), user stories, acceptance criteria
• ux.md - User flows, interactions, states, errors
• plan.md - Architecture, components, technology stack
• tasks.md - Implementation phases, TDD cycles, execution order

Data & Contracts

• data-model.md - Entities, validation, state machines
• contracts/ - APIs, messages, schemas
  • openapi.yaml
  • contracts.md

Supporting Documents

• research.md - Technical decisions, rationale
• setup.md - Environment, dependencies, commands

Step 3: Build Mental Model

Extract and internalize from each artifact:

From spec.md:

• Main journey and goal
• All requirements with IDs
• Test scenarios
• Edge cases and constraints

From ux.md:

• Complete flow with triggers
• States and transitions
• Error handling patterns

From plan.md:

• Code structure (A/​B/​C/​D patterns)
• Requirement→Component mapping
• Technical constraints

From tasks.md:

• Implementation sequence
• TDD coverage targets
• File creation paths
• Dependencies

From data-model.md:

• Field definitions and rules
• Constants and limits
• Entity relationships

From contracts:

• Endpoint specifications
• Message formats
• Protocol definitions

Step 4: Confirm Readiness

Output minimal confirmation:

✅ Feature context loaded: [feature-name]
   Available: [list of loaded artifacts]
   Ready for: implementation | checklist | queries

Error Handling

• Missing core files: Stop and report specific missing file
• Malformed content: Report file and parsing error
• Access denied: Report permission issue

Usage Notes

This skill prepares context for:

• Task execution (implementation from tasks.md)
• Checklist generation (requirements validation)
• Architecture queries
• Technical guidance

Context remains in memory for entire conversation.