Verify

Evidence-based local verification and explicitly requested Claude↔Codex crosschecks of improvement ideas, plans, and changes. Automatically selects independent proposals, draft critique, or mutual review, with candid recommendations and minimal exchanges. Shared by Claude Code and Codex. Generic verification stays local; discussing this skill is not permission to call another model. Direct multi-angle unit-test requests use the unit-test skill.

1 starsAI, LLMs & Agents Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add moveju112/skill_verify

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add moveju112/skill_verify --skill verify
Manual install
git clone --depth 1 https://github.com/moveju112/skill_verify.git /tmp/skill_verify
mkdir -p ~/.claude/skills
cp -r /tmp/skill_verify/skills/verify ~/.claude/skills/verify

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Verify

One shared entrypoint for Claude Code and Codex. The current agent remains the host and owns implementation; the counterpart may propose alternatives and review but never edits.

## Route and permission

- Ordinary verification: inspect and test locally; never call an external LLM.
- Crosscheck: require explicit task-scoped permission to involve the selected external reviewer. Loading this skill, installing it, or asking for generic verification is not permission.
- Claude host → Codex reviewer. Codex host → Claude reviewer; never invoke Codex recursively from Codex.
- Respect an explicitly requested reviewer, but do not describe same-family review as independent cross-model verification; clarify that conflict before calling.
- Infer the stopping stage separately from the strategy below: `analyze` for opinions or exploration, `plan` for a plan, `verify` for completed work, and `full` only when implementation is genuinely requested and authorized. An ambiguous advice request never defaults to `full`.
- Review-only requests stay read-only. Ask separately before DB, network, remote, destructive, deployment, Git mutation, or service-changing operations beyond the authorized reviewer call.

## Automatic crosscheck strategy

Choose from the existing question, draft, constraints, and impact; do not ask the user to choose a strategy. Respect an explicitly requested method. Automatic routing does not grant external-model permission.
Read the full skill on GitHub

Details

Skill name
verify
Path
skills/verify/SKILL.md
License
MIT
Repo updated
5 October 2026

FAQ

What does the Verify skill do?

Evidence-based local verification and explicitly requested Claude↔Codex crosschecks of improvement ideas, plans, and changes. Automatically selects independent proposals, draft critique, or mutual review, with candid recommendations and minimal exchanges. Shared by Claude Code and Codex. Generic verification stays local; discussing this skill is not permission to call another model. Direct multi-angle unit-test requests use the unit-test skill.

How do I install the Verify skill?

Claude Code plugin: /plugin marketplace add moveju112/skill_verify. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Verify popular?

It lives in moveju112/skill_verify, which has 1 GitHub stars and 0 forks, and is included in 1 curated skill list.

Related skills

Crosscheck

1

moveju112/skill_verify

Claude-hosted Claude↔Codex crosschecks of improvement ideas, plans, and completed changes. Automatically selects independent proposals, draft critique, or mutual review and allows candid rejection. Use for explicit requests such as "크로스체크", "코덱스랑 핑퐁", "둘이 의견 취합해", "cross-check with codex", or "/crosscheck". Discussing or editing this skill does not authorize another model call.

AI, LLMs & AgentsIn 1 list

Unit Test

1

moveju112/skill_verify

Use when the user directly requests efficiency-aware multi-angle unit tests or a test report after code changes. Triggers include "단위 테스트", "유닛테스트", "테스트 돌려", "다방면 테스트", "테스트 문서 정리", and the English equivalents "unit test", "run the tests", "multi-angle tests", "write up the test results", and "/unit-test". Codex or Claude runs tests locally without calling another LLM, avoids duplicate full-suite gates, and writes a focused report to the project's test directory. Not for E2E, external DB/network checks, runtime verification, or performance benchmarks.

Testing & QAIn 1 list

Client Workspace

1

CalebAnderson-Coder/ai-agency-os-starter

Create and maintain an isolated workspace per client, with a client CLAUDE.md holding brand voice, rules, banned claims, contacts, access map, and logs. Use when adding a client, switching clients, or when Claude must "act as" a specific client.

AI, LLMs & AgentsIn 1 list

Close Kitchen

5

TalonT-Org/AutoSkillit

Close the AutoSkillit kitchen — hides kitchen MCP tools for this session.

AI, LLMs & Agents

Prompt Engineering

5

Dreydrey9000/claude-skills

Expert guide on prompt engineering patterns, best practices, and optimization techniques. Use when user wants to improve prompts, learn prompting strategies, or debug agent behavior.

AI, LLMs & Agents

Open Kitchen

5

TalonT-Org/AutoSkillit

Open the AutoSkillit kitchen — reveals all kitchen MCP tools for this session. Human-only entry point.

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.