Feature Extraction Workflow

Extract atomic features from VISION.md into FEATURES.md.

Context

Skill to load: Load the katachi:framework-core skill for workflow principles.

Required files: @planning/VISION.md - Project vision to extract features from @planning/FEATURES.md - Current feature inventory (if exists)

General Guidance

Follow the collaborative workflow principles from the framework-core skill.

Features-specific guidance:

Use a scratchpad - Track state in /tmp/features-state.md:

• Raw features identified with source traceability
• Agent validation findings
• Refinements to apply
• Category decisions

Extract all at once - Review entire vision and extract all features in one pass.

Detect gaps proactively - Challenge completeness:

• "What handles errors here?"
• "What validates this input?"
• "Should we consider security/logging/configuration?"

Process

0. Check Existing State

If planning/FEATURES.md exists:

• Read current feature inventory
• Read current vision (check for changes)
• Compare: Are there new workflows? Changed scope?
• Ask: "Should we review for new features, or refine existing ones?"
• Enter iteration mode as appropriate

If no features exist: proceed with initial extraction

1. Extract All Features at Once

Review the vision document thoroughly. Extract all features in a single pass:

Consider all aspects:

• Core workflows and discrete capabilities
• Foundational capabilities
• Infrastructure features
• Platform integration
• Cross-cutting concerns (security, logging, config, errors)
• User feedback mechanisms

For each feature, document source traceability (which vision section).

2. Agent Validation #1: Raw Feature List

Dispatch a general-purpose subagent to review the raw feature list.

Provide: vision document, raw feature list

Request critique on:

• Completeness: All workflows covered? What's missing?
• Redundancies: Duplicate features?
• Atomicity: Single-session features? Split or merge needed?
• Gaps: What workflows lack supporting features?

Review findings with user.

3. User Iteration on Raw Features

Based on agent validation:

• Split features that are too large
• Merge or remove redundant features
• Add missing features identified
• Ensure each feature is atomic and traceable

4. Analyze and Categorize

Propose 3-7 natural groupings based on the features.

For each category:

• Present all features in that category together
• Show: proposed ID, description, complexity for each
• User reviews the full category
• Validate/adjust each feature
• Move to next category

5. Agent Validation #2: Categorized Features

Dispatch a general-purpose subagent to review the completed inventory.

Request critique on:

• Completeness: All requirements covered?
• Atomicity: Single-session features?
• Category alignment: Features in right categories?
• Complexity assessment: Ratings reasonable?
• Gaps and redundancies: Remaining issues?
• Dependencies: Implicit dependency concerns?

Review findings with user.

6. User Iteration and Finalization

Ask: "Should we iterate based on this feedback, or is the inventory complete?"

If gaps to address → refine features If complete → finalize and write to planning/FEATURES.md

Atomicity Check

Apply to each feature:

• Can this be implemented in a single focused session?
• Does it do ONE thing?
• If no → split into smaller features

Workflow

This is a collaborative process:

• Extract features systematically
• Challenge gaps and completeness
• User confirms categories, complexity, IDs
• Challenge non-atomic features
• Never add features without user agreement
• Iterate until complete