requirements-definition

📁 kentoshimizu/sw-agent-skills 📅 1 day ago
0
总安装量
1
周安装量
安装命令
npx skills add https://github.com/kentoshimizu/sw-agent-skills --skill requirements-definition

Agent 安装分布

amp 1
cline 1
opencode 1
cursor 1
continue 1
kimi-cli 1

Skill 文档

Requirements Definition

Overview

Use this skill to convert validated evidence into a stable requirement baseline with traceability and ownership.

Scope Boundaries

  • Use this skill when the task matches the trigger condition described in description.
  • Do not use this skill when the primary task falls outside this skill’s domain.

Shared Requirements Contract

  • Governance contract:
    • references/requirements-governance-contract.md
  • Manifest field guide:
    • references/manifest-field-guide.md
  • Optional validator (if repository enforces manifest validation):
    • python3 scripts/validate_requirements_contract.py --manifest <path/to/manifest.json>

Templates And Assets

  • Requirement baseline checklist:
    • assets/requirements-baseline-checklist.md
  • Valid sample manifests:
    • assets/rqm-def-manifest.valid.json
    • assets/rqm-cmp-manifest.valid.json

Inputs To Gather

  • Elicitation evidence with source references.
  • Product goals, constraints, and operating assumptions.
  • Architecture and compliance constraints.

Deliverables

  • Canonical REQ-* baseline.
  • Scope and out-of-scope decisions.
  • Traceability map (requirement -> evidence).
  • Decision/open-issue log with owners.

Workflow

  1. Translate evidence into atomic requirement statements.
  2. Link every requirement to supporting evidence.
  3. Separate functional requirements and quality constraints.
  4. Validate baseline using assets/requirements-baseline-checklist.md.
  5. Publish baseline and unresolved decision items.

Quality Standard

  • Each requirement is atomic, testable, and source-traceable.
  • Scope boundaries are explicit.
  • Ownership and validation method are defined.

Failure Conditions

  • Stop when conflicting requirements lack decision authority.
  • Stop when mandatory constraints are not represented.
  • Escalate when baseline cannot be approved due to unresolved contradictions.