haojing8312/workclaw-change-verification

WorkClaw Change Verification

Overview

Use this skill to decide what "verified" means for the changed area in WorkClaw. Reuse the repo's existing commands and run the smallest command set that honestly covers the touched surface.

When to Use

• React runtime code changed
• Tauri backend code changed
• Sidecar runtime or adapters changed
• Rust package logic changed
• Builtin skill assets or skill-related crates changed
• Build or test behavior changed
• UI flows changed in a way that can affect user behavior

Do not use this for docs-only changes that cannot affect runtime, tests, packaging, or build behavior.

Command Routing

• Use pnpm test:sidecar for apps/runtime/sidecar/ changes.
• Use pnpm test:rust-fast for apps/runtime/src-tauri/ or packages/* Rust logic changes.
• Use pnpm test:builtin-skills for builtin skill asset or skill crate changes.
• Use pnpm test:e2e:runtime for user-facing runtime flows that need end-to-end coverage.
• Use pnpm build:runtime when desktop packaging, build output, or integration boundaries may be affected.

When multiple surfaces changed, combine the relevant commands instead of picking only one.

Required Output

• Commands run
• Pass or fail results
• Which changed areas were covered by each command
• Any areas still unverified
• Whether it is valid to claim the work is verified

Output Template

Use this shape:

## Verification Summary
- Changed surface:
- Commands run:
- Results:
- Covered areas:
- Still unverified:
- Verification verdict:

Core Pattern

1. Map changed files to runtime surface areas.
2. Select the smallest honest command set.
3. Run the commands before claiming completion.
4. Report verification concretely, not vaguely.

Common Mistakes

• Saying "verified" after reading tests without running them.
• Running only frontend tests for a change that also affects Tauri or sidecar behavior.
• Forgetting pnpm test:builtin-skills when skill assets or skill-related crates changed.