agentsclimarketplace
Sign in

Nodeflow validate

Skill CAI5946/NodeFlow/skills/nodeflow-validate

Use after implementation tasks to run project checks, verify core paths, record failures, and write docs/test-report.md.From its SKILL.md

Install
npx -y skills add CAI5946/NodeFlow --skill nodeflow-validate

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.4 KB, as published. Nobody here has run it

Nodeflow Validate

Node Metadata

  • Node ID: nodeflow-validate
  • 作用: 执行项目级验证并生成测试报告
  • 输入: AGENTS.md, docs/TASKS.md, 项目代码、可选 docs/CHANGELOG.md
  • 输出: docs/test-report.md
  • 前置条件: 实现任务已完成或需要项目级验证
  • Gate: 验证命令无法确定、关键依赖不可用或关键检查失败时停止
  • 默认 checkpoint: N
  • 可跳过条件: Low 风险任务已由 nodeflow-verify 覆盖且不需要报告
  • 下游节点: nodeflow-project-closeout, nodeflow-handoff
  • 适用风险: Medium / High

Node Status

  • Status: core
  • Role: validation node
  • Replacement: none

Core Rules

  • Reply in Chinese.
  • Validate observable behavior; do not add new features.
  • Use commands from AGENTS.md and verification commands from docs/TASKS.md.
  • Read command output before summarizing results.
  • Record failures honestly.
  • Do not claim the project is ready if critical checks fail.

Inputs

Required:

  • AGENTS.md
  • docs/TASKS.md
  • project code

Recommended if present:

  • docs/CHANGELOG.md
  • docs/ASSUMPTIONS.md
  • docs/ARCH.md
  • docs/CONTEXT.md

Actions

  1. Read validation commands from AGENTS.md.
  2. Read task verification evidence from docs/TASKS.md and docs/CHANGELOG.md.
  3. Run the project-level checks that are appropriate for the repo:
    • tests
    • lint/analyze
    • build
    • smoke run
  4. Check for:
    • startup failure
    • obvious runtime errors
    • core flow unavailable
    • file structure drift
    • duplicate implementation
    • unresolved assumptions
    • multi-agent interface conflict
  5. Generate docs/test-report.md.

test-report.md Structure

# Test Report

## 运行命令

## 测试结果

## 失败项

## 已知问题

## 未解决假设

## 修复建议

## 是否可以进入收尾阶段

Outputs

  • docs/test-report.md

Completion Criteria

  • docs/test-report.md exists.
  • all run commands are recorded.
  • results include pass/fail status.
  • failures and known issues are explicit.
  • readiness for closeout is stated.

Stop Conditions

Stop and ask if:

  • required validation commands cannot be determined.
  • credentials, devices, or external services are required and unavailable.
  • critical tests fail and the user asked not to fix them in this stage.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.