OpenParable
Skip to content
← Back to skills

Prd To Features

Documents
prd-to-features

Prd To Features is an Agent Skill for documents published in the majiayu000/claude-skill-registry repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Convert PRD documents into features.json format with BDD test cases. Use this skill when users need to: (1) convert PRD/design documents into testable feature lists, (2) generate BDD test cases for PRD, (3) create features.json files.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/majiayu000/claude-skill-registry/blob/HEAD/skills/product/prd-to-features-baqif2-synapse-agent/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/prd-to-features, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

PRD to Features.json

Convert PRD documents into priority-ordered {prd_basename}-{number}.json files containing verifiable BDD test cases.

Conversion Workflow

Phase 1: Analyze PRD Completeness

  1. Read the PRD document
  2. Run the 6-dimension validation from completeness-checklist.md for each feature section
  3. List incomplete points and ask the user:
    • I'll provide details — User provides specific information
    • Use default assumptions — Mark in expected field with (Default assumption: ...)

Phase 2: Split into Verifiable Behaviors

After determining feature granularity, split each PRD section into independently verifiable behaviors:

PRD Section → Multiple features
Each feature → 1-N BDD test cases
Each BDD → Specific steps + Single expected result

Granularity Principle: Each BDD test case should independently verify a specific behavior.

Phase 3: Priority Ordering

Order features from high to low priority (array order represents priority):

  1. Core Infrastructure — Base components that other features depend on
  2. Core Routing/Dispatch — Core logic that determines system behavior
  3. Basic Capability Implementation — Concrete functionality implementation layer
  4. Tools/Converters — Supporting toolchain
  5. Discovery/Integration Mechanisms — Extensibility-related features
  6. User Interface/Experience — User-facing interaction features

Phase 4: Generate JSON

Generate features.json following the spec.md format:

{
  "category": "functional",
  "prd": "filename.md#section-anchor",
  "description": "Feature description",
  "bdd": [
    {
      "title": "Test title",
      "description": "Test description",
      "steps": ["Step 1", "Step 2"],
      "expected": "Expected result"
    }
  ],
  "passes": false
}

Phase 5: Check File Size

  • Lines ≤ 1000: Output single {prd_basename}-{number}.json
  • Lines > 1000: Split by feature module boundaries into prd-features-01.json, prd-features-02.json

Output Location

Default output to plans/ subdirectory at the same level as the PRD:

prd-directory/
├── 02-architecture.md    # PRD file
└── prd-features-01.json     # Generated file

Example Dialogue

User: Convert this PRD to features.json

Assistant:

  1. Read PRD, run completeness check
  2. List incomplete points (if any)
  3. Ask user how to handle them
  4. Confirm granularity and priority
  5. Generate features.json to plans/ directory