360 Secure SkillHub
360 安全技能中心

FDE PRD Writer

面向 FDE 交付流程第三阶段的 PRD 撰写 Skill:基于上游问题发现包与 POC 参与章程,锁定问题与目标、设计最小闭环、编写带编号的需求与验收标准、补充下游交接信息,并借助参考文档和质量清单检查可实现性、可测试性与可追溯性。默认输出英文 POC PRD 规范交接包。

xukun0821
xukun0821
热度 15.41
最近更新 2026-08-07 13:58:47
分类:开发依赖:代码执行 / 读写文件应用场景:报告撰写 / 模板 / 工作流

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add fde-prd-writer
Metadata
name
fde-prd-writer
description
Stage 3 of FDE Delivery Loop. Turn an approved problem-discovery package and POC engagement charter into an English POC PRD specification handoff for engineering, QA, deployment, Agent Skill design, and POC execution. Use to write, improve, or review POC PRDs, functional requirements, acceptance criteria, test scenarios, scope decisions, and downstream handoffs. Do not use for initial discovery, customer commitments, prototype construction, or production operations.

FDE PRD Writer | POC Specification Handoff

Turn an aligned customer problem and POC agreement into a PRD package that engineering can implement, QA can accept, and downstream teams can continue delivering.

Confirm upstream readiness

Before starting, confirm the presence of:

  1. The Customer Problem-Discovery Package from fde-problem-discovery.
  2. The POC Engagement Charter from fde-engagement-charter.

When either is missing, state the gap and return to the corresponding upstream stage. Do not fill the PRD with unvalidated assumptions.

Writing process

  1. Lock the problem and outcome: Extract target users, real workflow, business outcomes, success criteria, and explicit non-goals.
  2. Design the minimum closed loop: Define scope, functions, exceptions, permissions, data, and dependencies around user tasks. Replace vague terms such as “intelligent” and “easy to use” with observable behavior.
  3. Write testable specifications: Number each requirement. Define trigger, preconditions, main flow, exception flow, acceptance criteria, and evidence.
  4. Complete downstream handoffs: Identify the inputs and open risks required by Deployment Architecture, Agent Skill Design, POC Run, and Adoption and Value.
  5. Run quality checks: Use references/prd-quality.md. Rewrite or mark as open anything that cannot be implemented, tested, or traced.

Method selection principles

Treat KANO, SWOT, JTBD, user stories, and Mermaid as decision and communication tools:

  • Use a method only when it improves a scope, priority, risk, or workflow decision.
  • Convert every conclusion back into concrete evidence, requirements, and acceptance criteria.
  • Use Mermaid to clarify complex flows, states, or system relationships, not as decoration.

See references/method-selection.md for method selection and references/upstream-downstream-contracts.md for lifecycle boundaries.

Load references/user-stories.md, references/test-scenarios.md, andreferences/prd-style.md only when needed. Interviewing, discovery, and product strategy belong to Problem Discovery or Productization and are not the default starting point for this skill.

Output

Use templates/prd-skeleton.md to produce the POC PRD Specification Handoff Package. Default to English in this edition. Produce another language only when the user explicitly requests it.

Before delivery, run node scripts/validate-traceability.js <prd.md>to check structural traceability and orphan identifiers across FR → AC → TS. The script detects identifier and relationship gaps only; it cannot decide whether a requirement, acceptance criterion, or test is correct for the business.

Boundary

This skill does not replace Problem Discovery, the POC Charter, Deployment Architecture, Agent implementation, real execution, or value review. Its responsibility is to give those downstream activities a reliable, traceable, and acceptable specification.