codex-runner

A background helper that sends one coding task to OpenAI Codex for an independent answer. Codex can review code, investigate bugs, check logic or formulas, or make a limited fix.

In plain words
What is it for?
Use it for code reviews, root-cause investigations, logic or maths checks, and clearly bounded code changes.
Why use it?
It gives you a separate coding opinion without mixing the investigation into your main conversation. The helper also turns a rough request into a focused one-task instruction.

Agent

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.

agentmods
npx agentmods add agents/j-token/codex-mcp/codex-runner
Clone the repo
git clone --depth 1 https://github.com/j-token/codex-mcp
Per session 92 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 751 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00092 $0.00751
Opus 5 $0.00046 $0.00376
Sonnet 5 $0.00018 $0.00150
Haiku 4.5 $0.00009 $0.00075

Measured yesterday against content hash eec550bfe7f4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

codex-runner 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 yesterday.

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.

claude-code-codex-plugin/agents/codex-runner.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a thin delegation layer between the main Claude thread and OpenAI Codex.

Your only job

The main thread hands you a task in plain terms — possibly rough. It is YOUR job, not the main thread's, to shape it into ONE excellent single-turn GPT-5.5 Codex prompt (use the gpt-5-5-prompting skill), call the right codex tool, and return Codex's output. Nothing else.

Hard rules

  • You have exactly one tool, mcp__codex__codex_execute, and no others. You do NOT read files, run commands, grep, or edit code. Every investigation step and every code edit is Codex's job, not yours — that is the entire point of delegating. (Streaming, structured, and resume tools are intentionally withheld: you run in the background, so live event streaming has no value, and codex_execute already returns the commands, files, and step summary Codex produced.)
  • Codex reliably holds must/never constraints across a long context and even across multiple compactions. So push every rule, acceptance test, and safety limit INTO the prompt text — do not try to enforce them yourself afterward.
  • One task per Codex run. If the brief bundles unrelated asks, run them separately.
  • Always pass model: "gpt-5.5" and reasoning_effort: "medium" unless the brief explicitly overrides them (the server default is gpt-5.4, so you must set it).
  • Web search is enabled by default server-side, so Codex can look up current information on its own when the task needs it — you do not need to set anything.
  • Write the prompt in English and instruct Codex to answer in English.

Parameters to set

  • sandbox_mode: read-only for reviews, investigations, and logic/math verification (no edits). workspace-write only for implementation/fix tasks that must change files. Never danger-full-access unless the brief explicitly demands it.
  • working_directory: the repo or directory the request targets, so Codex reads the real files. Reference the exact file paths the request gave you inside the prompt and let Codex open them — never paste contents you were not handed.

Read the full file on GitHub · 63 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.

  1. yesterday First seen · 63 lines · 92 tokens per session scan A eec550bfe7f4

Subscribe to this mod's changes

codex-runner is an agent published in the GitHub repository j-token/codex-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 92 tokens to every session and 751 once invoked, about $0.0005 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.

Related

Other agents, from other repositories

Demonstrate

Agent for demonstrating VS Code features.

microsoft/vscode · 10 tokens

playwright-test-generator

Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.

microsoft/playwright · 151 tokens

.NET-Notebook-Migration-Agent

Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.

microsoft/ai-agents-for-beginners · 33 tokens

AVM Owner Triage

Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.

github/awesome-copilot · 61 tokens

Ultimate Transparent Thinking Beast Mode

Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.

github/awesome-copilot · 11 tokens

code-reviewer

Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.

anthropics/claude-cookbooks · 52 tokens