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/lifeshoftcode/codex-chatgpt-skill-recipe/chatgptnpx skills add lifeshoftcode/codex-chatgpt-skill-recipe --skill chatgptgit clone --depth 1 https://github.com/lifeshoftcode/codex-chatgpt-skill-recipeWrote 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/lifeshoftcode/codex-chatgpt-skill-recipe/chatgpt)<a href="https://agentmods.dev/skills/lifeshoftcode/codex-chatgpt-skill-recipe/chatgpt"><img src="https://agentmods.dev/badge/skills/lifeshoftcode/codex-chatgpt-skill-recipe/chatgpt.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 | $0.00095 | $0.02386 |
| Opus 5 | $0.00048 | $0.01193 |
| Sonnet 5 | $0.00019 | $0.00477 |
| Haiku 4.5 | $0.00010 | $0.00239 |
Grade A, and why
chatgpt 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ChatGPT Browser Bridge
Use this skill as a bridge from Codex to the ChatGPT website through the Codex in-app Browser.
Two modes are supported:
- Open/view mode: put
https://chatgpt.com/in front of the user. - Delegated consultation mode: when the user clearly asks Codex to consult, talk with, validate with, or ask ChatGPT about the current work, Codex may write and submit the prompt to ChatGPT, wait for the full response to finish, then read the answer and bring the useful result back to the user.
Behavior
- Use the Browser / in-app browser capability to navigate to
https://chatgpt.com/. - When authentication is required, assume the user already has an authenticated ChatGPT session in the Codex in-app Browser. Otherwise, ask the user to sign in there and retry before submitting prompts.
- Reuse an existing ChatGPT tab when practical, but begin a new conversation for a new delegated consultation unless the user explicitly asks to continue an existing consultation.
- Keep the browser visible when the user asks to open or view ChatGPT.
- After navigation, verify that the tab is not blank. A useful check is the current URL plus a light page observation such as title, DOM snapshot, or screenshot.
- If Browser reports
net::ERR_ABORTED, the tab title remainsabout:blank/New tab, the screenshot stays blank, a CAPTCHA or security check blocks access, ChatGPT is rate-limited, the requested model is unavailable, or the connection is lost, report the concrete blocker and stop. Offer Chrome or the official API as a later fallback only if helpful. - Do not use
web.run, the OS default browser, Chrome, or a raw shellstartcommand for this workflow. - Do not upload files, paste secrets, credentials, tokens, payment data, private family/student data, or large raw repo dumps into ChatGPT. User authorization to send context does not override these privacy and secret-handling rules.
Conversation Isolation
For delegated consultations, start a new ChatGPT conversation unless the user explicitly asks to continue an existing consultation.
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.
- 4d ago First seen · 131 lines · 95 tokens per session scan A c2b199fb0c95
chatgpt is a skill published in the GitHub repository lifeshoftcode/codex-chatgpt-skill-recipe (2 stars, last pushed 2mo ago), licensed MIT. It adds 95 tokens to every session and 2,386 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.
Other skills, from other repositories
color-expert
Use when working with color naming, color theory, color spaces, color definitions, or any task involving color knowledge - palettes, ramps, gradients, conversions, accessibility, perceptual matching, pigment mixing, print-vs-screen color, CSS color syntax, or historical color terminology. Use this skill whenever the…
read-image
当用户要求查看、读取、描述或分析图片/截图时使用（当前主模型不支持图像输入）。把图片交给配置的纯视觉模型识别，并把识别结果转述给用户。支持截图、报错图、图表、扫描件、照片等。.
codex-imagegen
Use when the user asks to generate, create, draw, edit, or restyle an image, or when a task needs image assets — logos, icons, illustrations, hero images, textures, photos, placeholder art, social/OG cards. Most coding agents cannot produce raster images; a Codex session via the codex CLI has a built-in imagegen tool…
codex-swarm
Use BEFORE writing or editing code, tests, docs, or other content yourself — whenever the writing spans more than one file or more than 20 lines. Delegating the writing to Codex sessions via the codex CLI is the default; you orchestrate. Also use when the user asks to swarm/fan out/parallelize work across Codex or…
project-brain-bridge
Use when creating, deploying, diagnosing, or operating a reusable ChatGPT-as-project-brain bridge for any local Codex project: initialize .ai-bridge state, expose read-only project evidence, expose a narrow stage-plan write connector, avoid stale/freeform connector schemas, and run the ChatGPT plan -> Codex execution…
dcf-model
Build discounted cash flow valuation workbooks in Excel.