tapd-story-implement

Executes TDD-based code implementation for story tasks via routed developer subagents.

840|261|Updated Apr 18, 2019
One-click install
npx skills add https://github.com/TencentBlueKing/bk-bcs --skill tapd-story-implement
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: tapd-story-implement
Source: https://github.com/TencentBlueKing/bk-bcs/tree/main/.agents/skills/tapd-story-pipeline/tapd-story-implement
Command: npx skills add https://github.com/TencentBlueKing/bk-bcs --skill tapd-story-implement

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Coordinating the code implementation phase of a story pipeline is error-prone: tasks must be executed in TDD mode, backend and frontend work must be routed to the right agents, and failures need controlled retry or rollback without corrupting the working branch. This Skill standardizes that implementation stage.

Core Features & Use Cases

  • Agent Routing by tech_type: Routes implementation to backend-developer, frontend-developer, or a serial two-round fullstack flow based on meta.yaml.
  • TDD Implementation via speckit: Invokes /speckit.implement against tasks.md so all tasks are completed test-first with code-scope whitelist enforcement.
  • Controlled Retry and Rollback: Supports in-place round retries (up to 3) and attempt-level re-entry with path-scoped git checkout to the baseline commit, excluding the work directory.
  • Use Case: A confirmed story with spec.md, plan.md, and tasks.md is implemented by dispatching the backend developer first, then the frontend developer against the plan's API contract, with phase advancing to implemented only after tests pass.

Quick Start

Run the implement stage for the current story using the existing spec.md, plan.md, and tasks.md in the work directory.

Frequently Asked Questions about tapd-story-implement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run the implementation phase of a story pipeline?

Ensure the story phase is confirmed and spec.md, plan.md, and tasks.md exist in the work directory. The pipeline records a baseline commit, then dispatches the appropriate developer subagent to execute tasks.md in TDD mode via speckit.implement.

How does fullstack implementation routing work?

Fullstack stories run two serial rounds: backend-developer first, then frontend-developer. The frontend round treats the plan.md API contract as authoritative and uses the backend result only as context. Progress is tracked in meta.yaml implement_rounds.

What happens when an implementation attempt fails?

On failure you choose retry, rollback re-entry, or termination. Retries fix code in place without git rollback for up to 3 rounds; rollback re-entry checks out only code-scope paths to the baseline commit, excluding the work directory.

Does this skill create or switch git branches?

No. It assumes you are already on the iteration branch and never creates or switches branches. The only git mutation allowed is path-scoped checkout to the baseline commit during cross-attempt rollback.

When does implementation roll back to earlier pipeline phases?

Unclear interface contracts roll back to specify, missing modules or data models roll back to plan, and omitted task scenarios roll back to tasks. The target phase is recorded in meta.yaml last_failure before exit.