Hermes Codex Imagegen Skill
This skill allows you to generate images directly within Hermes Agent using the Codex CLI. If you have a paid subscription to ChatGPT, you can generate images via the Codex CLI…
Install
git clone https://github.com/madrobotnet/hermes-codex-imagegen-skillInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →hermes-codex-imagegen-skill
ENGLISH | 한국어
English
Overview
A Hermes Agent skill for CLI-based image generation through a separate Codex CLI process authenticated with ChatGPT / OpenAI Codex OAuth.
This skill exists for the case where Hermes' native image backend is unavailable, but codex is installed and already logged in.
TL;DR
- a CLI-only fallback for Hermes users who already have a working Codex login
- runs
codex execin a separate process instead of touching Hermes' native image backend - verifies the real PNG output from
~/.codex/generated_images/ - useful when you want a practical Codex-based imagegen workaround, not an API-native integration
Why this is CLI-based
This workflow depends on:
- the local
codexexecutable codex login statuscodex exec --skip-git-repo-check --ephemeral ...- verifying generated PNG files under
~/.codex/generated_images/
So this is not a generic OpenAI Images API wrapper. It is a Codex CLI workaround workflow packaged as a Hermes skill.
What it does
- launches a separate Codex CLI instance
- asks Codex to generate an image using its built-in
imagegenskill - captures the Codex session id
- re-verifies the actual output file from the Hermes side
- warns that
bwrapor sandbox inspection errors do not always mean the image generation failed
What it does not do
- it does not rewire Hermes' native
image_generatetool - it does not directly call the OpenAI Images API
- it does not remove the need for a working Codex login
Requirements
- Hermes Agent
- Codex CLI installed and on
PATH - active ChatGPT / OpenAI Codex OAuth login
- a Linux or macOS shell environment where
codex execcan run
Repository contents
SKILL.md— the Hermes skill itselfREADME.md— public overviewLICENSE— MIT license.gitignore— minimal repo hygiene
Published names
- Repository name:
hermes-codex-imagegen-skill - Skill slug:
codex-imagegen-via-chatgpt-oauth
The repository name stays short and readable. The skill slug stays more explicit because the auth path matters for execution.
Quick publish flow
git init
git add .
git commit -m "Prepare the Hermes Codex image generation skill for publication"
gh repo create hermes-codex-imagegen-skill --public --source . --push
Notes
Generated image files normally land under:
~/.codex/generated_images/<session_id>/...
If you want to use a generated file in a real project, copy it into the project after verification instead of leaving it only under .codex.
↑ Back to top
한국어
개요
이 저장소는 Hermes Agent용 스킬이며, CLI 기반으로 동작합니다.
핵심 아이디어는 ChatGPT / OpenAI Codex OAuth로 로그인된 별도 Codex CLI 프로세스를 이용해 이미지를 생성하는 것입니다.
Hermes의 기본 이미지 백엔드를 바로 사용하기 어려운 상황에서, 로컬에 codex가 설치되어 있고 이미 로그인까지 완료되어 있을 때 사용할 수 있는 우회 워크플로우입니다.
TL;DR
- Codex 로그인이 이미 되어 있으면 바로 활용할 수 있는 CLI 전용 fallback입니다.
- Hermes 기본 이미지 백엔드는 건드리지 않고, 별도
codex exec프로세스로 우회합니다. - 실제 결과물은
~/.codex/generated_images/에서 다시 검증합니다. - OpenAI Images API 통합이라기보다, 실용적인 Codex 기반 imagegen workaround에 가깝습니다.
왜 CLI 기반인가요?
이 워크플로우는 다음 전제에 의존합니다.
- 로컬
codex실행 파일 codex login statuscodex exec --skip-git-repo-check --ephemeral ...~/.codex/generated_images/아래에 저장된 PNG 결과 검증
즉, 이것은 일반적인 OpenAI Images API 래퍼가 아닙니다.
Codex CLI를 활용하는 이미지 생성 우회 워크플로우를 Hermes 스킬 형태로 정리한 것입니다.
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…