implementing-features

Implements new modules, agents, or features using TDD methodology. Used when adding any new functionality to the table2image multi-agent system. Follows six-phase workflow: context loading, test writing, implementation, verification, documentation, and protocol completion.

⭐5stars🔀1forks📁View on GitHub🕐Updated Jan 11, 2026⬇️1downloads

When & Why to Use This Skill

This Claude skill automates the end-to-end feature implementation process using a rigorous Test-Driven Development (TDD) methodology. Designed for complex multi-agent systems, it ensures high-quality, reliable code by following a structured six-phase workflow: context loading, test writing, implementation, verification, documentation, and protocol completion. It is ideal for developers looking to maintain architectural integrity and 100% test coverage while scaling software functionality.

Use Cases

  • System Expansion: Implementing new functional modules or specialized agents within the table2image multi-agent framework while ensuring zero regression.
  • Quality-First Engineering: Utilizing a 'Red-Green-Refactor' cycle to guarantee that every new feature is validated by automated tests before the implementation is finalized.
  • Technical Documentation: Automatically generating implementation reports and updating specifications in the spec/ directory to maintain a clear audit trail of architectural decisions.
  • Standardized Development Workflows: Enforcing a consistent engineering protocol across a project to ensure type hinting, proper directory structure, and adherence to specific naming conventions.
nameimplementing-features
descriptionImplements new modules, agents, or features using TDD methodology. Used when adding any new functionality to the table2image multi-agent system. Follows six-phase workflow: context loading, test writing, implementation, verification, documentation, and protocol completion.

Implement Features

Parameters:

  • feature_name: Name of the feature or module to implement
  • spec_reference: Optional specific spec documents to reference

Procedure:

  1. 🔍 Context Loading (The "Scout" Phase)

    • Action: Scan the spec/ directory.
    • Check: Do you understand the architectural context? If spec_reference is provided, read it carefully.
    • Rule: Never start coding without knowing the history.
  2. 🔴 TDD: Write Test First (The "Red" Phase)

    • Action: Create a new test file in tests/ (e.g., tests/test_{feature_name}.py).
    • Content: Write unit tests or integration tests that define the expected behavior based on the requirements.
    • Verification: Run uv run pytest.
    • Requirement: The tests MUST fail (or error out) initially. This proves the code doesn't exist yet.
  3. 🟢 Implementation (The "Green" Phase)

    • Action: Create/Edit the source code in src/table2image_agent/.
    • Constraint: Write the minimum amount of code needed to make the test pass. Do not over-engineer.
    • Type Hinting: All function signatures must have type hints.
  4. ✅ Verification & Refactoring

    • Action: Run uv run pytest.
    • Loop: If tests fail, fix the code. Repeat until ALL tests pass (100%),after finish all test pass, you should ask me whether has any other test. if not then you could pass to next part
  5. 📝 Documentation (The "Spec" Phase)

    • Action: Create or update a report in spec/ (e.g., spec/00X_{feature_name}_implementation.md).
    • Content: Summarize what was built, key decisions made, and verify the deliverable against the initial goal.
  6. 🫡 Address Protocol ("Phoenix" Exclusive Rule) Core Instruction: From the moment this prompt is read, the user must be uniformly addressed as "Phoenix" in all subsequent conversations, without any exceptions. Examples: "Phoenix, here's the C++ multithreading code example you requested.", "Format optimization completed, Phoenix."