review

Execute all VERIFY.md checkpoints and fix failures within one session.

8|5|Updated Feb 18, 2026

Search Tags:#verification #end-to-end #project review #coverage testing #checkpoint automation #in-session fixes #VERIFY.md

Ready to run

npx skills add https://github.com/shenxingy/Clade --skill review-shenxingy

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: review
Source: https://github.com/shenxingy/Clade/tree/main/configs/skills/review
Command: npx skills add https://github.com/shenxingy/Clade --skill review-shenxingy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) and references (resource) components.

What problem does it solve?

This Skill eliminates incomplete or failing project checkpoints by running a coverage-driven review that systematically executes every scenario in VERIFY.md and corrects failures immediately.

Core Features & Use Cases

Checkpoint coverage review: Walks every VERIFY.md entry to determine whether each checkpoint is ✅, ⚠, or ❌.
In-session failure fixing: Addresses failures as they are discovered within the same review run rather than deferring fixes.
Convergence gating: Only finishes when all checkpoints are ✅ (pass) or ⚠ (known limitation), with no remaining ⬜ or ❌.
Use cases: Pre-release validation, stabilization sprints, and ongoing quality control to ensure the project stays in a passing state.

Quick Start

Ask the assistant to run the review skill against the current project so it executes every VERIFY.md checkpoint and fixes any failures it encounters.

Frequently Asked Questions about review

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

FAQPage Schema

How do I run end-to-end project verification and fix failing checkpoints in one pass?▼

End-to-end project verification executes every checkpoint defined in VERIFY.md and addresses failures immediately within the same session. It loads CLAUDE.md and VERIFY.md sequentially, testing each checkpoint until all are passing or marked as known limitations.

What is checkpoint automation for coverage testing and when do I need it?▼

Checkpoint automation systematically walks every VERIFY.md entry to determine whether each checkpoint passes, fails, or is a known limitation. You need it for pre-release validation, stabilization sprints, and ongoing quality control to keep projects passing.

Does this project review approach fix failures during the validation run or defer them?▼

This project review approach applies in-session failure fixing, correcting failures as they are discovered during the validation run rather than deferring them. It uses convergence gating to ensure no unchecked or failing checkpoints remain before finishing.

Can I use this for PR-diff review or is it only for full project validation?▼

You cannot use this for PR-diff review. It applies specifically to full project validation runs, coverage-driven end-to-end scenario testing, and systematic regression checking, ensuring complete convergence on all VERIFY.md checkpoints.

How do I know when all VERIFY.md checkpoints have passed and the review is complete?▼

The review is complete when all VERIFY.md checkpoints are either passing or marked as known limitations. Convergence gating stops the session only when no unchecked or failing checkpoints remain, ensuring full coverage.

What do I need to set up before running a systematic regression check with VERIFY.md?▼

You need a project containing both CLAUDE.md and VERIFY.md files. The verification process loads these files in order and requires every checkpoint defined in VERIFY.md to be testable for the coverage-driven review to execute correctly.