check-work

Verify completed work by spawning a subagent that reviews diffs, runs builds, and tests.

Updated May 28, 2026
One-click install
npx skills add https://github.com/changfengpro/agent-skills --skill check-work-changfengpro
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: check-work
Source: https://github.com/changfengpro/agent-skills/tree/main/skills/check-work
Command: npx skills add https://github.com/changfengpro/agent-skills --skill check-work-changfengpro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After completing a coding or operational task, it is easy to miss bugs, incomplete requirements, or broken builds. This Skill adds an independent verification pass that reviews the session's work, runs builds and tests, and reports a clear PASS or FAIL verdict before you consider the task done. ## Core Features & Use Cases - Subagent-Based Verification: Spawns a general-purpose verifier subagent that independently reviews the conversation trace, inspects the environment, and validates outcomes rather than trusting claims. - Two-Phase Review: Always performs a trace review of what was asked versus what was done, and runs a code review phase (diff inspection, build, tests, linters) whenever code is involved. - Iterative Fix Loop: On a FAIL verdict, the identified issues are fixed and verification repeats up to three times until the work passes. - Use Case: After implementing an authentication fix, invoke the skill with a focus area like "auth logic and JWT handling" to have the verifier confirm the fix compiles, passes tests, and fully addresses the request. ## Quick Start Ask the assistant to run /check-work after finishing a task to verify the changes with an independent review.

Frequently Asked Questions about check-work

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify my code changes with a subagent?

Invoke /check-work after completing your task. It spawns a general-purpose verifier subagent that reviews the conversation trace, inspects the git diff, runs builds and tests, and returns a PASS or FAIL verdict with specific issues.

How to verify only a specific part of my changes?

Pass a focus area when invoking the skill, such as /check-work auth logic and JWT handling. The verifier scopes its verdict to that area while still using the full session context to understand what was requested and done.

What happens when the verification verdict is FAIL?

On a FAIL verdict, the skill fixes the issues the verifier identified and runs verification again, repeating up to three times. Each FAIL report includes exact error output, file paths, line numbers, and suggested fixes.

Does the verifier check non-code tasks like deployments or research?

Yes. The trace review phase covers operational tasks, git and PR actions, research, and Q&A. The verifier confirms outcomes directly, such as checking that a submitted job was accepted or that a created PR exists, rather than trusting claims.

What causes an automatic FAIL during verification?

A broken build or failing tests cause an automatic FAIL. Violations of rules explicitly stated in the repository's AGENTS.md or Claude.md files also cause a FAIL, with the rule and file location cited in the report.