From gpt-image
Use when you want GPT Image 2 generation through Hermes native OpenAI Codex auth instead of OPENAI_API_KEY. This adapts the upstream gpt_image_2_skill workflow for Hermes by using ChatGPT/Codex OAuth and the image_generation tool path.
How this skill is triggered — by the user, by Claude, or both
Slash command
/gpt-image:gpt-image-codex-hermesThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Hermes-native adaptation of the upstream `wuyoscar/gpt_image_2_skill` project.
agents/openai.yamlreferences/craft.mdreferences/gallery-anime-and-manga.mdreferences/gallery-architecture-and-interior.mdreferences/gallery-beauty-and-lifestyle.mdreferences/gallery-brand-systems-and-identity.mdreferences/gallery-character-design.mdreferences/gallery-cinematic-and-animation.mdreferences/gallery-cinematic-film-references.mdreferences/gallery-data-visualization.mdreferences/gallery-edit-endpoint-showcase.mdreferences/gallery-events-and-experience.mdreferences/gallery-fashion-editorial.mdreferences/gallery-fine-art-painting.mdreferences/gallery-gaming.mdreferences/gallery-illustration.mdreferences/gallery-infographics-and-field-guides.mdreferences/gallery-ink-and-chinese.mdreferences/gallery-isometric.mdreferences/gallery-more-illustration-styles.mdHermes-native adaptation of the upstream wuyoscar/gpt_image_2_skill project.
The upstream project assumes direct OpenAI API access via OPENAI_API_KEY. This adaptation replaces that with Hermes's existing openai-codex auth flow, so image generation runs through ChatGPT/Codex OAuth and does not require a separate OpenAI API key.
gpt-image-2 in Hermes without OPENAI_API_KEYDo not use this when:
images.edit / mask editing todayConfirm Hermes has Codex auth:
hermes auth list openai-codex
Recommended Hermes config:
hermes config set image_gen.provider openai-codex
hermes config set image_gen.model gpt-image-2-medium
hermes config set image_gen.openai-codex.model gpt-image-2-medium
uv run ~/.hermes/skills/creative/gpt-image-codex-hermes/scripts/generate.py -p "PROMPT"
If you cloned this repo locally instead of installing the skill into Hermes:
uv run skills/gpt-image-codex-hermes/scripts/generate.py -p "PROMPT"
Examples:
uv run skills/gpt-image-codex-hermes/scripts/generate.py \
-p "cinematic cyberpunk street in Tokyo rain, dense neon signage" \
--size landscape --quality medium
uv run skills/gpt-image-codex-hermes/scripts/generate.py \
-p 'tea poster with exact Chinese text "山川茶事" and elegant editorial layout' \
--size portrait --quality high -f ./poster.png
Supported now:
Not supported yet:
This repo preserves the upstream gallery/reference material under skills/gpt-image-codex-hermes/references/ so prompt examples are still useful even though the auth/runtime path changed.
hermes auth list openai-codex shows an available credentialimage_gen.provider is set to openai-codex in Hermes configOPENAI_API_KEYGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub felipeorlando/gpt-image-codex-hermes --plugin gpt-image