Back to Details

feature-implement

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Feature Implement Skill

功能实现Skill

Purpose

目的

Execute the implementation plan by dispatching tasks to domain agents wave by wave. Phase 3 of the feature lifecycle (design → plan → implement → validate → release).
通过分阶段向领域Agent分配任务来执行实现计划。属于功能生命周期的第3阶段（设计→计划→实现→验证→发布）。

Operator Context

操作者上下文

Hardcoded Behaviors (Always Apply)

硬编码行为（始终适用）

  • CLAUDE.md Compliance: Read and follow repository CLAUDE.md
  • Plan Required: CANNOT implement without a plan in 
    .feature/state/plan/
  • State Management via Script: All state operations through 
    python3 ~/.claude/scripts/feature-state.py
  • Domain Agent Dispatch: Every task dispatched to its assigned domain agent via Task tool
  • Wave Order Enforcement: Complete all tasks in Wave N before starting Wave N+1
  • Wave Checkpoint: Run relevant tests after each wave completes
  • Deviation Handling: Tier 1-2 auto-fix, Tier 3 stops for user
  • 遵循CLAUDE.md规范：阅读并遵循仓库中的CLAUDE.md
  • 必须有计划：在
    .feature/state/plan/
    目录下无计划时，无法执行实现操作
  • 通过脚本管理状态：所有状态操作需通过
    python3 ~/.claude/scripts/feature-state.py
    执行
  • 分配领域Agent：每个任务需通过Task工具分配给对应的领域Agent
  • 强制分阶段执行：完成第N阶段的所有任务后，才能开始第N+1阶段
  • 阶段检查点：每个阶段完成后运行相关测试
  • 偏差处理：1-2级偏差自动修复，3级偏差需暂停并交由用户处理

Default Behaviors (ON unless disabled)

默认行为（默认开启，可关闭）

  • Context Loading: Read L0, L1, plan artifact, and design artifact at prime
  • Fresh Agent Per Task: Each task gets a clean agent dispatch (no context pollution)
  • Spec Compliance Check: After each task, verify output matches plan specification
  • Progress Reporting: Report after each task and wave completion
  • 上下文加载：初始化时读取L0、L1、计划工件和设计工件
  • 任务独立Agent：每个任务分配全新的Agent，避免上下文污染
  • 规范合规检查：每个任务完成后，验证输出是否符合计划规范
  • 进度报告：每个任务和阶段完成后进行报告

Optional Behaviors (OFF unless enabled)

可选行为（默认关闭，可开启）

  • Parallel within wave: Dispatch parallel-safe tasks simultaneously
  • Auto-fix Tier 2 deviations: Handle missing dependencies automatically
  • 阶段内并行执行：同时分配支持并行的任务
  • 自动修复2级偏差：自动处理依赖缺失等问题

What This Skill CAN Do

本Skill可实现的功能

  • Dispatch tasks to domain agents (golang-general-engineer, typescript-frontend-engineer, etc.)
  • Execute wave-ordered plans with dependency tracking
  • Handle deviations with tiered escalation
  • Run wave checkpoints (tests) between waves
  • 向领域Agent（golang-general-engineer、typescript-frontend-engineer等）分配任务
  • 执行带依赖追踪的分阶段计划
  • 按层级处理偏差
  • 在阶段之间运行检查点测试

What This Skill CANNOT Do

本Skill不可实现的功能

  • Implement without a plan
  • Override domain agent selection from plan
  • Skip wave ordering
  • Handle Tier 3 (architectural) deviations without user input
  • 无计划时执行实现操作
  • 覆盖计划中指定的领域Agent选择
  • 跳过阶段顺序
  • 在无用户输入的情况下处理3级（架构级）偏差

Instructions

操作说明

Phase 0: PRIME

阶段0：初始化

  1. Verify feature state:
    bash
    python3 ~/.claude/scripts/feature-state.py status FEATURE
    Verify current phase is 
    implement
    and 
    plan
    is completed.
  2. Load plan artifact from 
    .feature/state/plan/
    .
  3. Load design artifact from 
    .feature/state/design/
    for reference.
  4. Load L1 implement context:
    bash
    python3 ~/.claude/scripts/feature-state.py context-read FEATURE L1 --phase implement
  5. Capture BASE_SHA:
    bash
    git rev-parse HEAD
Gate: Plan loaded. Feature in implement phase. BASE_SHA captured. Proceed.
  1. 验证功能状态：
    bash
    python3 ~/.claude/scripts/feature-state.py status FEATURE
    验证当前阶段为
    implement
    且
    plan
    已完成。
  2. 从
    .feature/state/plan/
    加载计划工件。
  3. 从
    .feature/state/design/
    加载设计工件作为参考。
  4. 加载L1实现上下文：
    bash
    python3 ~/.claude/scripts/feature-state.py context-read FEATURE L1 --phase implement
  5. 捕获BASE_SHA：
    bash
    git rev-parse HEAD
检查点：已加载计划，功能处于实现阶段，已捕获BASE_SHA。继续执行。

Phase 1: EXECUTE (Wave Dispatch)

阶段1：执行（阶段任务分配）

For each wave in the plan:
Step 1: Dispatch Tasks
For each task in the wave:
  1. Check if task is parallel-safe and parallel mode is enabled
  2. Dispatch to assigned domain agent via Task tool:
Agent(
  subagent_type="[agent-from-plan]",
  prompt="Implement the following task:\n\n[TASK DETAILS FROM PLAN]\n\nFiles: [FILES]\nOperations: [OPERATIONS]\nVerification: [VERIFICATION COMMAND]\n\nContext from design: [RELEVANT DESIGN CONTEXT]",
  description="Implement [task title]"
)
  1. Verify task output matches plan specification
Step 2: Handle Deviations
TierExamplesAction
Tier 1: Auto-FixBug, type error, missing importAuto-apply, record in retro
Tier 2: BlockingMissing dependency, config issueAuto-fix if possible, record
Tier 3: ArchitecturalSchema change, API change, scope expansionSTOP, present to user
Check gate for Tier 3: 
python3 ~/.claude/scripts/feature-state.py gate FEATURE implement.architectural-deviation
Step 3: Wave Checkpoint
After all tasks in a wave complete:
  1. Run the project's test suite (or relevant subset)
  2. If tests fail: identify which task caused the failure, route back to that agent
  3. If tests pass: proceed to next wave
Step 4: Progress Report
After each wave:
Wave N complete: X/Y tasks passed
  T1: [agent] ✓
  T2: [agent] ✓
  T3: [agent] ✗ (deviation: [description])
Gate: All waves complete. All tasks verified. Proceed to Validate.
针对计划中的每个阶段：
步骤1：分配任务
针对阶段中的每个任务：
  1. 检查任务是否支持并行且已开启并行模式
  2. 通过Task工具分配给指定的领域Agent：
Agent(
  subagent_type="[agent-from-plan]",
  prompt="Implement the following task:\n\n[TASK DETAILS FROM PLAN]\n\nFiles: [FILES]\nOperations: [OPERATIONS]\nVerification: [VERIFICATION COMMAND]\n\nContext from design: [RELEVANT DESIGN CONTEXT]",
  description="Implement [task title]"
)
  1. 验证任务输出是否符合计划规范
步骤2：偏差处理
层级示例操作
1级：自动修复bug、类型错误、缺失导入自动修复，记录到回顾文档
2级：阻塞问题依赖缺失、配置问题尽可能自动修复，记录问题
3级：架构级schema变更、API变更、范围扩展暂停，提交给用户处理
检查3级偏差的检查点：
python3 ~/.claude/scripts/feature-state.py gate FEATURE implement.architectural-deviation
步骤3：阶段检查点
阶段内所有任务完成后：
  1. 运行项目的测试套件（或相关子集）
  2. 若测试失败：定位导致失败的任务，将其退回给对应的Agent修复
  3. 若测试通过：进入下一个阶段
步骤4：进度报告
每个阶段完成后：
第N阶段完成：X/Y个任务通过
  T1: [agent] ✓
  T2: [agent] ✓
  T3: [agent] ✗ (偏差：[描述])
检查点：所有阶段完成，所有任务验证通过。进入验证阶段。

Phase 2: VALIDATE (Implementation Review)

阶段2：验证（实现评审）

Quick validation before formal validation phase:
  • All planned files were created/modified
  • All verification commands pass
  • No unplanned files were modified (check 
    git diff --name-only BASE_SHA..HEAD
    )
  • No Tier 3 deviations unresolved
Gate: Implementation complete. Proceed to Checkpoint.
正式验证阶段前的快速验证：
  • 所有计划中的文件已创建/修改
  • 所有验证命令执行通过
  • 无计划外的文件被修改（检查
    git diff --name-only BASE_SHA..HEAD
    ）
  • 无未解决的3级偏差
检查点：实现完成。进入检查点阶段。

Phase 3: CHECKPOINT

阶段3：检查点

  1. Save implementation artifact (summary of what was built):
    bash
    echo "IMPL_SUMMARY" | python3 ~/.claude/scripts/feature-state.py checkpoint FEATURE implement
  2. Record learnings — if this phase produced non-obvious insights, record them:
    bash
    python3 ~/.claude/scripts/learning-db.py record TOPIC KEY "VALUE" --category design
  3. Advance:
    bash
    python3 ~/.claude/scripts/feature-state.py advance FEATURE
  4. Suggest next step:
    Implementation complete. Run /feature-validate for quality gates.
  1. 保存实现工件（已构建内容的摘要）：
    bash
    echo "IMPL_SUMMARY" | python3 ~/.claude/scripts/feature-state.py checkpoint FEATURE implement
  2. 记录经验 — 若本阶段产生了非显而易见的见解，记录下来：
    bash
    python3 ~/.claude/scripts/learning-db.py record TOPIC KEY "VALUE" --category design
  3. 推进阶段：
    bash
    python3 ~/.claude/scripts/feature-state.py advance FEATURE
  4. 建议下一步操作：
    实现已完成。运行/feature-validate进行质量检查。

Error Handling

错误处理

ErrorCauseSolution
No plan foundPlan phase not completedRun /feature-plan first
Agent dispatch failsAgent not available or task malformedRetry with more context, escalate if 3 failures
Wave test failureTask broke existing testsRoute back to responsible agent for fix
Tier 3 deviationArchitectural decision neededStop, present options to user
错误原因解决方案
未找到计划计划阶段未完成先运行/feature-plan
Agent分配失败Agent不可用或任务格式错误补充更多上下文后重试，若失败3次则升级处理
阶段测试失败任务破坏了现有测试将问题退回给负责的Agent修复
3级偏差需要架构决策暂停，向用户提供选项

Anti-Patterns

反模式

Anti-PatternWhy WrongDo Instead
Implement without dispatching to agentsBypasses domain expertiseUse Task tool to dispatch
Skip wave checkpointsFailures compound across wavesTest after every wave
Ignore deviationsSmall deviations become big problemsClassify and handle per tier
Dispatch all tasks in parallelFile conflicts cause corruptionRespect wave ordering and parallel-safe flags
反模式问题所在正确做法
不分配给Agent直接实现绕过了领域专业能力使用Task工具分配任务
跳过阶段检查点故障会在多个阶段累积每个阶段完成后都要测试
忽略偏差小偏差会演变成大问题按层级分类处理偏差
所有任务并行分配文件冲突会导致损坏遵循阶段顺序和并行安全标记

References

参考资料

  • Gate Enforcement
  • Retro Loop
  • State Conventions
  • Subagent-Driven Development
  • 检查点强制执行
  • 回顾循环
  • 状态约定
  • 子Agent驱动开发