Task Decomposition

技能: task-decomposition Purpose

Break a raw task into an ordered, executable, dependency-aware step 列出 with acceptance criteria and explicit open questions.

When to Use The 请求 implies multiple files or multiple sub系统s. The 请求 is vague, inconsistent, or underspecified. The 请求 mixes de签名 + implementation + 验证. 输入s raw_task_description (required, string): the user 请求 as-is. constrAInts (optional, string[]): non-negotiables (security, time, language, 工具ing). repo_上下文 (optional, string): relevant paths, conventions, or prior decisions. risk_level_hint (optional, enum: low|medium|high): if the user already 签名aled risk. Steps Re状态 the task in 1–3 sentences without 添加ing assumptions. 提取 deliverables (expected behavior, files to touch, commands to 运行). Identify unknowns that block execution and convert them into concrete questions. Split work into atomic steps; each step must include: action + tar获取 a single primary outcome acceptance criteria (done_when) Order steps by dependency and mark safe parallelization explicitly. Tag each step with: risk (low|medium|high) 验证 (what will be 检查ed) If unknowns are material, 停止 and ask only the minimum questions; otherwise proceed with 状态d assumptions. 验证 No step depends on hidden 上下文. Every step has measurable acceptance criteria. Dependencies are explicit (no “and then it works”). No step contAIns vague verbs (“improve”, “优化”, “make better”) without a measurable tar获取. 输出

Structured plan (example 模式):

summary: "" open_questions: - "" assumptions: - "" steps: - id: 1 action: "" tar获取s: [""] risk: low 验证: "<检查 to 运行>" done_when: ""

Safety Rules Do not invent requirements, APIs, or file paths. If a step can be destructive, require explicit confirmation in the plan. Prefer the smallest viable step sequence; avoid gold-plating. Example

输入:

raw_task_description: “添加 a 命令行工具 command to 导出 报告s.” constrAInts: ["No breaking changes", "Must include tests"]

输出 (excerpt):

summary: "添加 报告 导出 command and tests" open_questions: - "What 输出 格式化s are required (json/csv/pdf)?" steps: - id: 1 action: "Locate existing 命令行工具 entrypoints and command 路由r" tar获取s: ["src/命令行工具/*"] risk: low 验证: "命令行工具 help shows existing commands unchanged" done_when: "Command dis补丁 mechanism is identified"