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.
npx agentmods add skills/practicalswan/agent-skills/codex-routernpx skills add PracticalSwan/agent-skills --skill codex-routergit clone --depth 1 https://github.com/PracticalSwan/agent-skillsWrote 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.
[](https://agentmods.dev/skills/practicalswan/agent-skills/codex-router)<a href="https://agentmods.dev/skills/practicalswan/agent-skills/codex-router"><img src="https://agentmods.dev/badge/skills/practicalswan/agent-skills/codex-router.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00157 | $0.01546 |
| Opus 5 | $0.00078 | $0.00773 |
| Sonnet 5 | $0.00031 | $0.00309 |
| Haiku 4.5 | $0.00016 | $0.00155 |
Grade A, and why
codex-router 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.
How it starts
The opening of the file, as written. The whole thing — 129 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codex Router (custom models in the Codex app)
You are a custom model. The Codex app routes your traffic through codex-router.
How your tools work
- The app's native tools appear in your tool list with flattened names:
codex_app__create_thread,codex_app__list_threads,mcp__node_repl__js,mcp__peekaboo__create_task, and so on. - Call them with exactly those names. The router restores the original
namespace (for example
create_threadincodex_app) before the app sees the call, so the app executes it natively. - The router never executes an app tool. It only relays definitions and results. If a call fails, fix your arguments; do not try to run the tool yourself.
- Never spawn a side-channel driver. Do not start your own node_repl process, do not fake MCP metadata, do not write driver scripts. The tools you need are already in your tool list.
Before each kind of work, read the matching skill
- Threads, automations, navigation: read
codex-app-threads. - In-app browser: read
codex-in-app-browser. - Computer use: read
codex-computer-use.
When a tool rejects your arguments
The app answers received invalid arguments. when you missed a required
field. Stop guessing. Read the matching skill for the exact shape, then
retry once with the correct arguments. Repeated guessing burns tokens and
turns.
Golden rules
- Use the tools you were given. Do not build workarounds.
- Read the companion skill before the relevant work.
- When a call fails, fix the arguments from the skill, then retry.
- A turn with no tool call ends the task. After a tool result, if more work still needs a tool, call it in the same turn. Do not only announce the next step. Text-only is for when the user's request is fully done.
Spawned threads and model inheritance
For a new local Codex thread, omit the model field unless the user
explicitly requested one. The router selects the parent routed model. An
explicit model is never overridden. Follow-up messages retain the target
thread's settings, and cloud tasks choose their model outside this relay.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- yesterday Changed 00b848dc0220
- 5d ago First seen · 129 lines · 157 tokens per session scan A 4e374b066310
codex-router is a skill published in the GitHub repository PracticalSwan/agent-skills (13 stars, last pushed 4d ago), licensed MIT. It adds 157 tokens to every session and 1,546 once invoked, about $0.0008 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 2026-08-30.
Other skills, from other repositories
course-guide
Topic router for the AI Engineering from Scratch curriculum. Give it a topic, a question, or a bug you are fighting, and it points at the exact lessons that teach it, plus the right next command. Trigger phrases: "where do I learn", "which lesson covers", "course guide", "I'm stuck on", "what should I do next", "teach…
course-guide
AI Engineering from Scratch 课程的主题路由器。给它一个主题、问题或正在处理的 bug， 它会指出精确教授它的课程，以及下一条正确命令。触发短语： “在哪里学习”、“哪节课涵盖”、“课程导航”、“我卡在”、“接下来该做什么”、 “教我 MCP”、“教我 Agent Skills”、“在哪里准备 Claude certification”，或 "where do I learn", "which lesson covers", "course guide", "I'm stuck on", "what should I do next", "teach me MCP", "teach me Agent…
auto
Intent-classified router, the front door to OrchestKit and the DEFAULT entry point for any goal-shaped request. Classifies a plain-English goal and routes it to the right specialist skill. Routing is never overhead, so use it even when the target skill seems obvious; skip only when already executing inside another…
frontmcp-config
Use when configuring a FrontMCP server through frontmcp.config or the @FrontMcp options. Covers auth modes (public, transparent, local, remote), OAuth plus credential vault and secureStore, CORS, HTTP port / entry-path prefix / unix socket, security headers (CSP, HSTS, X-Frame-Options, X-Content-Type-Options), rate…
frontmcp-observability
Use when adding tracing, structured logging, metrics, or monitoring to a FrontMCP server. Covers zero-config OpenTelemetry distributed tracing across all flows; the this.telemetry API for custom spans, events, and attributes in tools, plugins, agents, and skills; structured JSON logging with trace correlation and…
a0-plugin-router
Main entry point for all Agent Zero plugin tasks. Routes to specialist skills for creating, reviewing, contributing, managing, or debugging plugins. Use when the user mentions plugins, asks how the plugin system works, wants to build/install/uninstall/publish/debug a plugin, or asks about the Plugin Hub.