What is multi-agent-verify?

A verification workflow that asks three separate coding agents to review, test, and potentially fix code in separate Git worktrees, which are isolated working copies of a repository. It then combines their findings and applies a final fix on the original branch.

Verifying code changes, running tests, finding bugs, and applying fixes before work is accepted or merged.

It reduces reliance on one review or test run when checking completed changes. Separate worktrees let the agents investigate independently without overwriting one another's files.

multi-agent-verify — Skill for Codex by leyuan0602-glitch

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add leyuan0602-glitch/cc-claude-codex --skill multi-agent-verify

Clone the repo

git clone --depth 1 https://github.com/leyuan0602-glitch/cc-claude-codex

Made for: Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify/github.svg)](https://agentmods.dev/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify)

Your own site

<a href="https://agentmods.dev/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify"><img src="https://agentmods.dev/badge/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify"><img src="https://agentmods.dev/badge/skills/leyuan0602-glitch/cc-claude-codex/multi-agent-verify.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 57 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,177 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00057	$0.02177
Opus 5	$0.00028	$0.01089
Sonnet 5	$0.00011	$0.00435
Haiku 4.5	$0.00006	$0.00218

Measured 9d ago against content hash 5c7932a4b5ce, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

multi-agent-verify scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

multi-agent-verify/SKILL.md · 212 lines

How it starts

The opening of the file, as written. The whole thing — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Multi-Agent Verify: Parallel Verification with 3 Independent Agents

You are the Main Agent (Claude Code). You orchestrate 3 independent agents to verify code in parallel. Each agent works in its own git worktree, reviews code, writes tests, runs verification, and fixes bugs independently. You then synthesize their findings and apply a final fix.

Agent Launch Methods

Each agent uses a DIFFERENT launch mechanism:

Agent	Method	Why
OpenCode	`Bash` CLI: `opencode run -m opencode/minimax-m2.5-free "..."`	External CLI, free model, no nesting issues
Codex	`Bash` CLI: `codex exec --full-auto "..."`	External CLI, no nesting issues
Claude	`Task` tool with `isolation: "worktree"`	MUST use Task tool — `claude` CLI CANNOT run nested inside Claude Code (detects `CLAUDECODE` env var and refuses)

CRITICAL: Never launch claude via Bash CLI from within Claude Code. Always use the Task tool with isolation: "worktree" and run_in_background: true.

Prerequisites

Before invoking:

All code changes are committed on the current branch
Requirements are available (from status.md or conversation context)
opencode and codex are in PATH (check with which)
No external orchestrator script is required — launch agents directly

Phase V1: Preparation

Mode A: cc-claude-codex Integration

When called as cc-claude-codex Phase 3:

Read .cc-claude-codex/status.md — extract ALL requirements and scenarios
Run git diff main...HEAD --name-status to get the full list of changed files
- If main doesn't exist, try master or use the appropriate base branch
Build the verification prompt (see Prompt Template below)
Write the filled prompt to .cc-claude-codex/verify-prompt-{ts}.md

Mode B: Standalone

When invoked directly:

Get requirements from conversation context, or ask the user
Identify changed files via git diff {base}...HEAD --name-status
- Try main, master, or ask user for the base branch
Build the verification prompt (see Prompt Template below)
Write the filled prompt to .claude/verify-prompt-{ts}.md

Read the full file on GitHub · 212 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

9d ago First seen · 212 lines · 57 tokens per session scan A 5c7932a4b5ce

Subscribe to this mod's changes

multi-agent-verify is a skill published in the GitHub repository leyuan0602-glitch/cc-claude-codex (9 stars, last pushed 6mo ago), licensed MIT. It adds 57 tokens to every session and 2,177 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Code review E2E testing Testing Unit testing

test-audit

Batch audit of test files against Q1-Q25 quality gates and AP1-AP32 anti-patterns. Detects orphan tests, phantom mocks, untested public methods. Tiered output (A/B/C/D) with critical gate enforcement and optional post-audit fix workflow. Flags: zuvo:test-audit all | [path] | [file] | --deep | --quick | --include-e2e |…

greglas75/zuvo · 100 tokens

QA Sentinel Agent

Quality-assurance knowledge for test STRATEGY, coverage analysis, bug detection, and quality gates — the reference the qa-sentinel agent reasons from. Use when designing a test plan, deciding unit vs integration vs E2E, judging whether coverage is real or hollow, setting quality-gate thresholds, or reviewing a PR for…

nxtg-ai/forge-plugin · 108 tokens

test-evidence-review

Quality review of test files and manual evidence docs. Goes beyond existence checks — evaluates assertion coverage, edge cases, naming conventions, and evidence completeness. Verdict: ADEQUATE / INCOMPLETE / MISSING per story. Run before QA sign-off or on demand.

cenconq25/claude-code-app-studio · 58 tokens

multi-agent-verify