{"page":{"pageid":1541,"slug":"skill-openai-system-imagegen","title":"imagegen skill (openai/skills)","content":"**What it does.** Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas. Part of [[skills-openai-skills]] (openai/skills).\n\n| | |\n| --- | --- |\n| Upstream | [openai/skills](https://github.com/openai/skills) |\n| Skill file | [skills/.system/imagegen/SKILL.md](https://github.com/openai/skills/blob/HEAD/skills/.system/imagegen/SKILL.md) |\n| License | Apache-2.0 (skill folder LICENSE.txt) |\n| Author | OpenAI |\n| Fetched | 2026-09-10 |\n\n## Install\n\n- Codex: `$skill-installer` installs from this catalog (this one ships with Codex by default); other agents: `npx skills add openai/skills --skill imagegen`.\n- Raw file: `curl -sL https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/SKILL.md`\n\n## SKILL.md (verbatim)\n\n```yaml\nname: \"imagegen\"\ndescription: \"Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas.\"\n```\n\n# Image Generation Skill\n\nGenerates or edits images for the current project (for example website assets, game assets, UI mockups, product mockups, wireframes, logo design, photorealistic images, or infographics).\n\n## Top-level modes and rules\n\nThis skill has exactly two top-level modes:\n\n- **Default built-in tool mode (preferred):** built-in `image_gen` tool for normal image generation and editing. Does not require `OPENAI_API_KEY`.\n- **Fallback CLI mode (explicit-only):** `scripts/image_gen.py` CLI. Use only when the user explicitly asks for the CLI path. Requires `OPENAI_API_KEY`.\n\nWithin the explicit CLI fallback only, the CLI exposes three subcommands:\n\n- `generate`\n- `edit`\n- `generate-batch`\n\nRules:\n- Use the built-in `image_gen` tool by default for all normal image generation and editing requests.\n- Never switch to CLI fallback automatically.\n- If the built-in tool fails or is unavailable, tell the user the CLI fallback exists and that it requires `OPENAI_API_KEY`. Proceed only if the user explicitly asks for that fallback.\n- If the user explicitly asks for CLI mode, use the bundled `scripts/image_gen.py` workflow. Do not create one-off SDK runners.\n- Never modify `scripts/image_gen.py`. If something is missing, ask the user before doing anything else.\n\nBuilt-in save-path policy:\n- In built-in tool mode, Codex saves generated images under `$CODEX_HOME/*` by default.\n- Do not describe or rely on OS temp as the default built-in destination.\n- Do not describe or rely on a destination-path argument (if any) on the built-in `image_gen` tool. If a specific location is needed, generate first and then move or copy the selected output from `$CODEX_HOME/generated_images/...`.\n- Save-path precedence in built-in mode:\n 1. If the user names a destination, move or copy the selected output there.\n 2. If the image is meant for the current project, move or copy the final selected image into the workspace before finishing.\n 3. If the image is only for preview or brainstorming, render it inline; the underlying file can remain at the default `$CODEX_HOME/*` path.\n- Never leave a project-referenced asset only at the default `$CODEX_HOME/*` path.\n- Do not overwrite an existing asset unless the user explicitly asked for replacement; otherwise create a sibling versioned filename such as `hero-v2.png` or `item-icon-edited.png`.\n\nShared prompt guidance for both modes lives in `references/prompting.md` and `references/sample-prompts.md`.\n\nFallback-only docs/resources for CLI mode:\n- `references/cli.md`\n- `references/image-api.md`\n- `references/codex-network.md`\n- `scripts/image_gen.py`\n\n## When to use\n- Generate a new image (concept art, product shot, cover, website hero)\n- Generate a new image using one or more reference images for style, composition, or mood\n- Edit an existing image (inpainting, lighting or weather transformations, background replacement, object removal, compositing, transparent background)\n- Produce many assets or variants for one task\n\n## When not to use\n- Extending or matching an existing SVG/vector icon set, logo system, or illustration library inside the repo\n- Creating simple shapes, diagrams, wireframes, or icons that are better produced directly in SVG, HTML/CSS, or canvas\n- Making a small project-local asset edit when the source file already exists in an editable native format\n- Any task where the user clearly wants deterministic code-native output instead of a generated bitmap\n\n## Decision tree\n\nThink about two separate questions:\n\n1. **Intent:** is this a new image or an edit of an existing image?\n2. **Execution strategy:** is this one asset or many assets/variants?\n\nIntent:\n- If the user wants to modify an existing image while preserving parts of it, treat the request as **edit**.\n- If the user provides images only as references for style, composition, mood, or subject guidance, treat the request as **generate**.\n- If the user provides no images, treat the request as **generate**.\n\nBuilt-in edit semantics:\n- Built-in edit mode is for images already visible in the conversation context, such as attached images or images generated earlier in the thread.\n- If the user wants to edit a local image file with the built-in tool, first load it with built-in `view_image` tool so the image is visible in the conversation context, then proceed with the built-in edit flow.\n- Do not promise arbitrary filesystem-path editing through the built-in tool.\n- If a local file still needs direct file-path control, masks, or other explicit CLI-only parameters, use the explicit CLI fallback only when the user asks for it.\n- For edits, preserve invariants aggressively and save non-destructively by default.\n\nExecution strategy:\n- In the built-in default path, produce many assets or variants by issuing one `image_gen` call per requested asset or variant.\n- In the explicit CLI fallback path, use the CLI `generate-batch` subcommand only when the user explicitly chose CLI mode and needs many prompts/assets.\n\nAssume the user wants a new image unless they clearly ask to change an existing one.\n\n## Workflow\n1. Decide the top-level mode: built-in by default, fallback CLI only if explicitly requested.\n2. Decide the intent: `generate` or `edit`.\n3. Decide whether the output is preview-only or meant to be consumed by the current project.\n4. Decide the execution strategy: single asset vs repeated built-in calls vs CLI `generate-batch`.\n5. Collect inputs up front: prompt(s), exact text (verbatim), constraints/avoid list, and any input images.\n6. For every input image, label its role explicitly:\n - reference image\n - edit target\n - supporting insert/style/compositing input\n7. If the edit target is only on the local filesystem and you are staying on the built-in path, inspect it with `view_image` first so the image is available in conversation context.\n8. If the user asked for a photo, illustration, sprite, product image, banner, or other explicitly raster-style asset, use `image_gen` rather than substituting SVG/HTML/CSS placeholders. If the request is for an icon, logo, or UI graphic that should match existing repo-native SVG/vector/code assets, prefer editing those directly instead.\n9. Augment the prompt based on specificity:\n - If the user's prompt is already specific and detailed, normalize it into a clear spec without adding creative requirements.\n - If the user's prompt is generic, add tasteful augmentation only when it materially improves output quality.\n10. Use the built-in `image_gen` tool by default.\n11. If the user explicitly chooses the CLI fallback, then and only then use the fallback-only docs for quality, `input_fidelity`, masks, output format, output paths, and network setup.\n12. Inspect outputs and validate: subject, style, composition, text accuracy, and invariants/avoid items.\n13. Iterate with a single targeted change, then re-check.\n14. For preview-only work, render the image inline; the underlying file may remain at the default `$CODEX_HOME/generated_images/...` path.\n15. For project-bound work, move or copy the selected artifact into the workspace and update any consuming code or references. Never leave a project-referenced asset only at the default `$CODEX_HOME/generated_images/...` path.\n16. For batches, persist only the selected finals in the workspace unless the user explicitly asked to keep discarded variants.\n17. Always report the final saved path for any workspace-bound asset, plus the final prompt and whether the built-in tool or fallback CLI mode was used.\n\n## Prompt augmentation\n\nReformat user prompts into a structured, production-oriented spec. Make the user's goal clearer and more actionable, but do not blindly add detail.\n\nTreat this as prompt-shaping guidance, not a closed schema. Use only the lines that help, and add a short extra labeled line when it materially improves clarity.\n\n### Specificity policy\n\nUse the user's prompt specificity to decide how much augmentation is appropriate:\n\n- If the prompt is already specific and detailed, preserve that specificity and only normalize/structure it.\n- If the prompt is generic, you may add tasteful augmentation when it will materially improve the result.\n\nAllowed augmentations:\n- composition or framing hints\n- polish level or intended-use hints\n- practical layout guidance\n- reasonable scene concreteness that supports the stated request\n\nNot allowed augmentations:\n- extra characters or objects that are not implied by the request\n- brand names, slogans, palettes, or narrative beats that are not implied\n- arbitrary side-specific placement unless the surrounding layout supports it\n\n## Use-case taxonomy (exact slugs)\n\nClassify each request into one of these buckets and keep the slug consistent across prompts and references.\n\nGenerate:\n- photorealistic-natural — candid/editorial lifestyle scenes with real texture and natural lighting.\n- product-mockup — product/packaging shots, catalog imagery, merch concepts.\n- ui-mockup — app/web interface mockups and wireframes; specify the desired fidelity.\n- infographic-diagram — diagrams/infographics with structured layout and text.\n- logo-brand — logo/mark exploration, vector-friendly.\n- illustration-story — comics, children’s book art, narrative scenes.\n- stylized-concept — style-driven concept art, 3D/stylized renders.\n- historical-scene — period-accurate/world-knowledge scenes.\n\nEdit:\n- text-localization — translate/replace in-image text, preserve layout.\n- identity-preserve — try-on, person-in-scene; lock face/body/pose.\n- precise-object-edit — remove/replace a specific element (including interior swaps).\n- lighting-weather — time-of-day/season/atmosphere changes only.\n- background-extraction — transparent background / clean cutout.\n- style-transfer — apply reference style while changing subject/scene.\n- compositing — multi-image insert/merge with matched lighting/perspective.\n- sketch-to-render — drawing/line art to photoreal render.\n\n## Shared prompt schema\n\nUse the following labeled spec as shared prompt scaffolding for both top-level modes:\n\n```text\nUse case: \nAsset type: \nPrimary request: \nInput images: (optional)\nScene/backdrop: \nSubject:
\nStyle/medium: \nComposition/framing: \nLighting/mood: \nColor palette: \nMaterials/textures: \nText (verbatim): \"\"\nConstraints: \nAvoid: \n```\n\nNotes:\n- `Asset type` and `Input images` are prompt scaffolding, not dedicated CLI flags.\n- `Scene/backdrop` refers to the visual setting. It is not the same as the fallback CLI `background` parameter, which controls output transparency behavior.\n- Fallback-only execution notes such as `Quality:`, `Input fidelity:`, masks, output format, and output paths belong in the explicit CLI path only. Do not treat them as built-in `image_gen` tool arguments.\n\nAugmentation rules:\n- Keep it short.\n- Add only the details needed to improve the prompt materially.\n- For edits, explicitly list invariants (`change only X; keep Y unchanged`).\n- If any critical detail is missing and blocks success, ask a question; otherwise proceed.\n\n## Examples\n\n### Generation example (hero image)\n```text\nUse case: product-mockup\nAsset type: landing page hero\nPrimary request: a minimal hero image of a ceramic coffee mug\nStyle/medium: clean product photography\nComposition/framing: wide composition with usable negative space for page copy if needed\nLighting/mood: soft studio lighting\nConstraints: no logos, no text, no watermark\n```\n\n### Edit example (invariants)\n```text\nUse case: precise-object-edit\nAsset type: product photo background replacement\nPrimary request: replace only the background with a warm sunset gradient\nConstraints: change only the background; keep the product and its edges unchanged; no text; no watermark\n```\n\n## Prompting best practices\n- Structure prompt as scene/backdrop -> subject -> details -> constraints.\n- Include intended use (ad, UI mock, infographic) to set the mode and polish level.\n- Use camera/composition language for photorealism.\n- Only use SVG/vector stand-ins when the user explicitly asked for vector output or a non-image placeholder.\n- Quote exact text and specify typography + placement.\n- For tricky words, spell them letter-by-letter and require verbatim rendering.\n- For multi-image inputs, reference images by index and describe how they should be used.\n- For edits, repeat invariants every iteration to reduce drift.\n- Iterate with single-change follow-ups.\n- If the prompt is generic, add only the extra detail that will materially help.\n- If the prompt is already detailed, normalize it instead of expanding it.\n- For explicit CLI fallback only, see `references/cli.md` and `references/image-api.md` for `quality`, `input_fidelity`, masks, output format, and output-path guidance.\n\nMore principles shared by both modes: `references/prompting.md`.\nCopy/paste specs shared by both modes: `references/sample-prompts.md`.\n\n## Guidance by asset type\nAsset-type templates (website assets, game assets, wireframes, logo) are consolidated in `references/sample-prompts.md`.\n\n## Fallback CLI mode only\n\n### Temp and output conventions\nThese conventions apply only to the explicit CLI fallback. They do not describe built-in `image_gen` output behavior.\n- Use `tmp/imagegen/` for intermediate files (for example JSONL batches); delete them when done.\n- Write final artifacts under `output/imagegen/`.\n- Use `--out` or `--out-dir` to control output paths; keep filenames stable and descriptive.\n\n### Dependencies\nPrefer `uv` for dependency management in this repo.\n\nRequired Python package:\n```bash\nuv pip install openai\n```\n\nOptional for downscaling only:\n```bash\nuv pip install pillow\n```\n\nPortability note:\n- If you are using the installed skill outside this repo, install dependencies into that environment with its package manager.\n- In uv-managed environments, `uv pip install ...` remains the preferred path.\n\n### Environment\n- `OPENAI_API_KEY` must be set for live API calls.\n- Do not ask the user for `OPENAI_API_KEY` when using the built-in `image_gen` tool.\n- Never ask the user to paste the full key in chat. Ask them to set it locally and confirm when ready.\n\nIf the key is missing, give the user these steps:\n1. Create an API key in the OpenAI platform UI: https://platform.openai.com/api-keys\n2. Set `OPENAI_API_KEY` as an environment variable in their system.\n3. Offer to guide them through setting the environment variable for their OS/shell if needed.\n\nIf installation is not possible in this environment, tell the user which dependency is missing and how to install it into their active environment.\n\n### Script-mode notes\n- CLI commands + examples: `references/cli.md`\n- API parameter quick reference: `references/image-api.md`\n- Network approvals / sandbox settings for CLI mode: `references/codex-network.md`\n\n## Reference map\n- `references/prompting.md`: shared prompting principles for both modes.\n- `references/sample-prompts.md`: shared copy/paste prompt recipes for both modes.\n- `references/cli.md`: fallback-only CLI usage via `scripts/image_gen.py`.\n- `references/image-api.md`: fallback-only API/CLI parameter reference.\n- `references/codex-network.md`: fallback-only network/sandbox troubleshooting for CLI mode.\n- `scripts/image_gen.py`: fallback-only CLI implementation. Do not load or use it unless the user explicitly chooses CLI mode.\n\n## Other files in this skill\n\n- [LICENSE.txt](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/LICENSE.txt)\n- [agents/openai.yaml](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/agents/openai.yaml)\n- [assets/imagegen-small.svg](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/assets/imagegen-small.svg)\n- [assets/imagegen.png](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/assets/imagegen.png)\n- [references/cli.md](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/references/cli.md)\n- [references/codex-network.md](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/references/codex-network.md)\n- [references/image-api.md](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/references/image-api.md)\n- [references/prompting.md](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/references/prompting.md)\n- [references/sample-prompts.md](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/references/sample-prompts.md)\n- [scripts/image_gen.py](https://raw.githubusercontent.com/openai/skills/HEAD/skills/.system/imagegen/scripts/image_gen.py)\n\n## references/cli.md (verbatim)\n\n# CLI reference (`scripts/image_gen.py`)\n\nThis file is for the fallback CLI mode only. Read it only after the user explicitly asks to use `scripts/image_gen.py` instead of the built-in `image_gen` tool.\n\n`generate-batch` is a CLI subcommand in this fallback path. It is not a top-level mode of the skill.\n\n## What this CLI does\n- `generate`: generate a new image from a prompt\n- `edit`: edit one or more existing images\n- `generate-batch`: run many generation jobs from a JSONL file\n\nReal API calls require **network access** + `OPENAI_API_KEY`. `--dry-run` does not.\n\n## Quick start (works from any repo)\nSet a stable path to the skill CLI (default `CODEX_HOME` is `~/.codex`):\n\n```\nexport CODEX_HOME=\"${CODEX_HOME:-$HOME/.codex}\"\nexport IMAGE_GEN=\"$CODEX_HOME/skills/.system/imagegen/scripts/image_gen.py\"\n```\n\nInstall dependencies into that environment with its package manager. In uv-managed environments, `uv pip install ...` remains the preferred path.\n\n## Quick start\n\nDry-run (no API call; no network required; does not require the `openai` package):\n\n```bash\npython \"$IMAGE_GEN\" generate \\\n --prompt \"Test\" \\\n --out output/imagegen/test.png \\\n --dry-run\n```\n\nNotes:\n- One-off dry-runs print the API payload and the computed output path(s).\n- Repo-local finals should live under `output/imagegen/`.\n\nGenerate (requires `OPENAI_API_KEY` + network):\n\n```bash\npython \"$IMAGE_GEN\" generate \\\n --prompt \"A cozy alpine cabin at dawn\" \\\n --size 1024x1024 \\\n --out output/imagegen/alpine-cabin.png\n```\n\nEdit:\n\n```bash\npython \"$IMAGE_GEN\" edit \\\n --image input.png \\\n --prompt \"Replace only the background with a warm sunset\" \\\n --out output/imagegen/sunset-edit.png\n```\n\n## Guardrails\n- Use the bundled CLI directly (`python \"$IMAGE_GEN\" ...`) after activating the correct environment.\n- Do **not** create one-off runners (for example `gen_images.py`) unless the user explicitly asks for a custom wrapper.\n- **Never modify** `scripts/image_gen.py`. If something is missing, ask the user before doing anything else.\n\n## Defaults\n- Model: `gpt-image-1.5`\n- Supported model family for this CLI: GPT Image models (`gpt-image-*`)\n- Size: `1024x1024`\n- Quality: `auto`\n- Output format: `png`\n- Default one-off output path: `output/imagegen/output.png`\n- Background: unspecified unless `--background` is set\n\n## Quality, input fidelity, and masks (CLI fallback only)\nThese are explicit CLI controls. They are not built-in `image_gen` tool arguments.\n\n- `--quality` works for `generate`, `edit`, and `generate-batch`: `low|medium|high|auto`\n- `--input-fidelity` is **edit-only** and validated as `low|high`\n- `--mask` is **edit-only**\n\nExample:\n\n```bash\npython \"$IMAGE_GEN\" edit \\\n --image input.png \\\n --prompt \"Change only the background\" \\\n --quality high \\\n --input-fidelity high \\\n --out output/imagegen/background-edit.png\n```\n\nMask notes:\n- For multi-image edits, pass repeated `--image` flags. Their order is meaningful, so describe each image by index and role in the prompt.\n- The CLI accepts a single `--mask`.\n- Use a PNG mask when possible; the script treats mask handling as best-effort and does not perform full preflight validation beyond file checks/warnings.\n- In the edit prompt, repeat invariants (`change only the background; keep the subject unchanged`) to reduce drift.\n\n## Output handling\n- Use `tmp/imagegen/` for temporary JSONL inputs or scratch files.\n- Use `output/imagegen/` for final outputs.\n- Reruns fail if a target file already exists unless you pass `--force`.\n- `--out-dir` changes one-off naming to `image_1.`, `image_2.`, and so on.\n- Downscaled copies use the default suffix `-web` unless you override it.\n\n## Common recipes\n\nGenerate with augmentation fields:\n\n```bash\npython \"$IMAGE_GEN\" generate \\\n --prompt \"A minimal hero image of a ceramic coffee mug\" \\\n --use-case \"product-mockup\" \\\n --style \"clean product photography\" \\\n --composition \"wide product shot with usable negative space for page copy\" \\\n --constraints \"no logos, no text\" \\\n --out output/imagegen/mug-hero.png\n```\n\nGenerate + also write a downscaled copy for fast web loading:\n\n```bash\npython \"$IMAGE_GEN\" generate \\\n --prompt \"A cozy alpine cabin at dawn\" \\\n --size 1024x1024 \\\n --downscale-max-dim 1024 \\\n --out output/imagegen/alpine-cabin.png\n```\n\nGenerate multiple prompts concurrently (async batch):\n\n```bash\nmkdir -p tmp/imagegen output/imagegen/batch\ncat > tmp/imagegen/prompts.jsonl << 'EOF'\n{\"prompt\":\"Cavernous hangar interior with a compact shuttle parked near the center\",\"use_case\":\"stylized-concept\",\"composition\":\"wide-angle, low-angle\",\"lighting\":\"volumetric light rays through drifting fog\",\"constraints\":\"no logos or trademarks; no watermark\",\"size\":\"1536x1024\"}\n{\"prompt\":\"Gray wolf in profile in a snowy forest\",\"use_case\":\"photorealistic-natural\",\"composition\":\"eye-level\",\"constraints\":\"no logos or trademarks; no watermark\",\"size\":\"1024x1024\"}\nEOF\n\npython \"$IMAGE_GEN\" generate-batch \\\n --input tmp/imagegen/prompts.jsonl \\\n --out-dir output/imagegen/batch \\\n --concurrency 5\n\nrm -f tmp/imagegen/prompts.jsonl\n```\n\nNotes:\n- `generate-batch` requires `--out-dir`.\n- generate-batch requires --out-dir.\n- Use `--concurrency` to control parallelism (default `5`).\n- Per-job overrides are supported in JSONL (for example `size`, `quality`, `background`, `output_format`, `output_compression`, `moderation`, `n`, `model`, `out`, and prompt-augmentation fields).\n- `--n` generates multiple variants for a single prompt; `generate-batch` is for many different prompts.\n- In batch mode, per-job `out` is treated as a filename under `--out-dir`.\n\n## CLI notes\n- Supported sizes: `1024x1024`, `1536x1024`, `1024x1536`, or `auto`.\n- Transparent backgrounds require `output_format` to be `png` or `webp`.\n- `--prompt-file`, `--output-compression`, `--moderation`, `--max-attempts`, `--fail-fast`, `--force`, and `--no-augment` are supported.\n- This CLI is intended for GPT Image models. Do not assume older non-GPT image-model behavior applies here.\n\n## See also\n- API parameter quick reference for fallback CLI mode: `references/image-api.md`\n- Prompt examples shared across both top-level modes: `references/sample-prompts.md`\n- Network/sandbox notes for fallback CLI mode: `references/codex-network.md`\n\n## references/codex-network.md (verbatim)\n\n# Codex network approvals / sandbox notes\n\nThis file is for the fallback CLI mode only. Read it only after the user explicitly asks to use `scripts/image_gen.py`.\n\nThis guidance is intentionally isolated from `SKILL.md` because it can vary by environment and may become stale. Prefer the defaults in your environment when in doubt.\n\n## Why am I asked to approve image generation calls?\nThe fallback CLI uses the OpenAI Image API, so it needs outbound network access. In many Codex setups, network access is disabled by default and/or the approval policy requires confirmation before networked commands run.\n\n## Important note about approvals vs network\n- `--ask-for-approval never` suppresses approval prompts.\n- It does **not** by itself enable network access.\n- In `workspace-write`, network access still depends on your Codex configuration (for example `[sandbox_workspace_write] network_access = true`).\n\n## How do I reduce repeated approval prompts?\nIf you trust the repo and want fewer prompts, use a configuration or profile that both:\n- enables network for the sandbox mode you plan to use\n- sets an approval policy that matches your risk tolerance\n\nExample `~/.codex/config.toml` pattern:\n\n```toml\napproval_policy = \"on-request\"\nsandbox_mode = \"workspace-write\"\n\n[sandbox_workspace_write]\nnetwork_access = true\n```\n\nIf you want quieter automation after network is enabled, you can choose a stricter approval policy, but do that intentionally and with care.\n\n## Safety note\nEnabling network and reducing approvals lowers friction, but increases risk if you run untrusted code or work in an untrusted repository.\n\n## references/image-api.md (verbatim)\n\n# Image API quick reference\n\nThis file is for the fallback CLI mode only. Use it only after the user explicitly asks to use `scripts/image_gen.py` instead of the built-in `image_gen` tool.\n\nThese parameters describe the Image API and bundled CLI fallback surface. Do not assume they are normal arguments on the built-in `image_gen` tool.\n\n## Scope\n- This fallback CLI is intended for GPT Image models (`gpt-image-1.5`, `gpt-image-1`, and `gpt-image-1-mini`).\n- The built-in `image_gen` tool and the fallback CLI do not expose the same controls.\n\n## Endpoints\n- Generate: `POST /v1/images/generations` (`client.images.generate(...)`)\n- Edit: `POST /v1/images/edits` (`client.images.edit(...)`)\n\n## Core parameters for GPT Image models\n- `prompt`: text prompt\n- `model`: image model\n- `n`: number of images (1-10)\n- `size`: `1024x1024`, `1536x1024`, `1024x1536`, or `auto`\n- `quality`: `low`, `medium`, `high`, or `auto`\n- `background`: output transparency behavior (`transparent`, `opaque`, or `auto`) for generated output; this is not the same thing as the prompt's visual scene/backdrop\n- `output_format`: `png` (default), `jpeg`, `webp`\n- `output_compression`: 0-100 (jpeg/webp only)\n- `moderation`: `auto` (default) or `low`\n\n## Edit-specific parameters\n- `image`: one or more input images. For GPT Image models, you can provide up to 16 images.\n- `mask`: optional mask image\n- `input_fidelity`: `low` (default) or `high`\n\nModel-specific note for `input_fidelity`:\n- `gpt-image-1` and `gpt-image-1-mini` preserve all input images, but the first image gets richer textures and finer details.\n- `gpt-image-1.5` preserves the first 5 input images with higher fidelity.\n\n## Output\n- `data[]` list with `b64_json` per image\n- The bundled `scripts/image_gen.py` CLI decodes `b64_json` and writes output files for you.\n\n## Limits and notes\n- Input images and masks must be under 50MB.\n- Use the edits endpoint when the user requests changes to an existing image.\n- Masking is prompt-guided; exact shapes are not guaranteed.\n- Large sizes and high quality increase latency and cost.\n- High `input_fidelity` can materially increase input token usage.\n- If a request fails because a specific option is unsupported by the selected GPT Image model, retry manually without that option.\n\n## Important boundary\n- `quality`, `input_fidelity`, explicit masks, `background`, `output_format`, and related parameters are fallback-only execution controls.\n- Do not assume they are built-in `image_gen` tool arguments.\n\n## references/prompting.md (verbatim)\n\n# Prompting best practices\n\nThese prompting principles are shared by both top-level modes of the skill:\n- built-in `image_gen` tool (default)\n- explicit `scripts/image_gen.py` CLI fallback\n\nThis file is about prompt structure, specificity, and iteration. Fallback-only execution controls such as `quality`, `input_fidelity`, masks, output format, and output paths live in the fallback docs.\n\n## Contents\n- [Structure](#structure)\n- [Specificity policy](#specificity-policy)\n- [Allowed and disallowed augmentation](#allowed-and-disallowed-augmentation)\n- [Composition and layout](#composition-and-layout)\n- [Constraints and invariants](#constraints-and-invariants)\n- [Text in images](#text-in-images)\n- [Input images and references](#input-images-and-references)\n- [Iterate deliberately](#iterate-deliberately)\n- [Fallback-only execution controls](#fallback-only-execution-controls)\n- [Use-case tips](#use-case-tips)\n- [Where to find copy/paste recipes](#where-to-find-copypaste-recipes)\n\n## Structure\n- Use a consistent order: scene/backdrop -> subject -> key details -> constraints -> output intent.\n- Include intended use (ad, UI mock, infographic) to set the level of polish.\n- For complex requests, use short labeled lines instead of one long paragraph.\n\n## Specificity policy\n- If the user prompt is already specific and detailed, normalize it into a clean spec without adding creative requirements.\n- If the prompt is generic, you may add tasteful detail when it materially improves the output.\n- Treat examples in `sample-prompts.md` as fully-authored recipes, not as the default amount of augmentation to add to every request.\n\n## Allowed and disallowed augmentation\n\nAllowed augmentation for generic prompts:\n- composition and framing cues\n- intended-use or polish-level hints\n- practical layout guidance\n- reasonable scene concreteness that supports the request\n\nDo not add:\n- extra characters, props, or objects that are not implied\n- brand palettes, slogans, or story beats that are not implied\n- arbitrary side-specific placement unless the surrounding layout supports it\n\n## Composition and layout\n- Specify framing and viewpoint (close-up, wide, top-down) and placement only when it materially helps.\n- Call out negative space if the asset clearly needs room for UI or copy.\n- Avoid making left/right layout decisions unless the user or surrounding layout supports them.\n\n## Constraints and invariants\n- State what must not change (`keep background unchanged`).\n- For edits, say `change only X; keep Y unchanged` and repeat invariants on every iteration to reduce drift.\n\n## Text in images\n- Put literal text in quotes or ALL CAPS and specify typography (font style, size, color, placement).\n- Spell uncommon words letter-by-letter if accuracy matters.\n- For in-image copy, require verbatim rendering and no extra characters.\n\n## Input images and references\n- Do not assume that every provided image is an edit target.\n- Label each image by index and role (`Image 1: edit target`, `Image 2: style reference`).\n- If the user provides images for style, composition, or mood guidance and does not ask to modify them, treat the request as generation with references.\n- If the user asks to preserve an existing image while changing specific parts, treat the request as an edit.\n- For compositing, describe how the images interact (`place the subject from Image 2 into Image 1`).\n\n## Iterate deliberately\n- Start with a clean base prompt, then make small single-change edits.\n- Re-specify critical constraints when you iterate.\n- Prefer one targeted follow-up at a time over rewriting the whole prompt.\n\n## Fallback-only execution controls\n- `quality`, `input_fidelity`, explicit masks, output format, and output paths are fallback-only execution controls.\n- Do not assume they are built-in `image_gen` tool arguments.\n- If the user explicitly chooses CLI fallback, see `references/cli.md` and `references/image-api.md` for those controls.\n\n## Use-case tips\nGenerate:\n- photorealistic-natural: Prompt as if a real photo is captured in the moment; use photography language (lens, lighting, framing); call for real texture; avoid over-stylized polish unless requested.\n- product-mockup: Describe the product/packaging and materials; ensure clean silhouette and label clarity; if in-image text is needed, require verbatim rendering and specify typography.\n- ui-mockup: Describe the target fidelity first (shippable mockup or low-fi wireframe), then focus on layout, hierarchy, and practical UI elements; avoid concept-art language.\n- infographic-diagram: Define the audience and layout flow; label parts explicitly; require verbatim text.\n- logo-brand: Keep it simple and scalable; ask for a strong silhouette and balanced negative space; avoid decorative flourishes unless requested.\n- illustration-story: Define panels or scene beats; keep each action concrete.\n- stylized-concept: Specify style cues, material finish, and rendering approach (3D, painterly, clay) without inventing new story elements.\n- historical-scene: State the location/date and required period accuracy; constrain clothing, props, and environment to match the era.\n\nEdit:\n- text-localization: Change only the text; preserve layout, typography, spacing, and hierarchy; no extra words or reflow unless needed.\n- identity-preserve: Lock identity (face, body, pose, hair, expression); change only the specified elements; match lighting and shadows.\n- precise-object-edit: Specify exactly what to remove/replace; preserve surrounding texture and lighting; keep everything else unchanged.\n- lighting-weather: Change only environmental conditions (light, shadows, atmosphere, precipitation); keep geometry, framing, and subject identity.\n- background-extraction: Request a clean cutout; crisp silhouette; no halos; preserve label text exactly; no restyling.\n- style-transfer: Specify style cues to preserve (palette, texture, brushwork) and what must change; add `no extra elements` to prevent drift.\n- compositing: Reference inputs by index; specify what moves where; match lighting, perspective, and scale; keep the base framing unchanged.\n- sketch-to-render: Preserve layout, proportions, and perspective; choose materials and lighting that support the supplied sketch without adding new elements.\n\n## Where to find copy/paste recipes\nFor copy/paste prompt specs (examples only), see `references/sample-prompts.md`. This file focuses on principles, specificity, and iteration patterns.\n\n## references/sample-prompts.md (verbatim)\n\n# Sample prompts (copy/paste)\n\nThese prompt recipes are shared across both top-level modes of the skill:\n- built-in `image_gen` tool (default)\n- explicit `scripts/image_gen.py` CLI fallback\n\nUse these as starting points. They are intentionally complete prompt recipes, not the default amount of augmentation to add to every user request.\n\nWhen adapting a user's prompt:\n- keep user-provided requirements\n- only add detail according to the specificity policy in `SKILL.md`\n- do not treat every example below as permission to invent extra story elements\n\nThe labeled lines are prompt scaffolding, not a closed schema. `Asset type` and `Input images` are prompt-only scaffolding; the CLI does not expose them as dedicated flags.\n\nExecution details such as explicit CLI flags, `quality`, `input_fidelity`, masks, output formats, and local output paths depend on mode. Use the built-in tool by default; only apply CLI-specific controls after the user explicitly opts into fallback mode.\n\nFor prompting principles (structure, specificity, invariants, iteration), see `references/prompting.md`.\n\n## Generate\n\n### photorealistic-natural\n```\nUse case: photorealistic-natural\nPrimary request: candid photo of an elderly sailor on a small fishing boat adjusting a net\nScene/backdrop: coastal water with soft haze\nSubject: weathered skin with wrinkles and sun texture\nStyle/medium: photorealistic candid photo\nComposition/framing: medium close-up, eye-level\nLighting/mood: soft coastal daylight, shallow depth of field, subtle film grain\nMaterials/textures: real skin texture, worn fabric, salt-worn wood\nConstraints: natural color balance; no heavy retouching; no glamorization; no watermark\nAvoid: studio polish; staged look\n```\n\n### product-mockup\n```\nUse case: product-mockup\nPrimary request: premium product photo of a matte black shampoo bottle with a minimal label\nScene/backdrop: clean studio gradient from light gray to white\nSubject: single bottle centered with subtle reflection\nStyle/medium: premium product photography\nComposition/framing: centered, slight three-quarter angle, generous padding\nLighting/mood: softbox lighting, clean highlights, controlled shadows\nMaterials/textures: matte plastic, crisp label printing\nConstraints: no logos or trademarks; no watermark\n```\n\n### ui-mockup\n```\nUse case: ui-mockup\nPrimary request: mobile app home screen for a local farmers market with vendors and daily specials\nAsset type: mobile app screen\nStyle/medium: realistic product UI, not concept art\nComposition/framing: clean vertical mobile layout with clear hierarchy\nConstraints: practical layout, clear typography, no logos or trademarks, no watermark\n```\n\n### infographic-diagram\n```\nUse case: infographic-diagram\nPrimary request: detailed infographic of an automatic coffee machine flow\nScene/backdrop: clean, light neutral background\nSubject: bean hopper -> grinder -> brew group -> boiler -> water tank -> drip tray\nStyle/medium: clean vector-like infographic with clear callouts and arrows\nComposition/framing: vertical poster layout, top-to-bottom flow\nText (verbatim): \"Bean Hopper\", \"Grinder\", \"Brew Group\", \"Boiler\", \"Water Tank\", \"Drip Tray\"\nConstraints: clear labels, strong contrast, no logos or trademarks, no watermark\n```\n\n### logo-brand\n```\nUse case: logo-brand\nPrimary request: original logo for \"Field & Flour\", a local bakery\nStyle/medium: vector logo mark; flat colors; minimal\nComposition/framing: single centered logo on a plain background with generous padding\nConstraints: strong silhouette, balanced negative space; original design only; no gradients unless essential; no trademarks; no watermark\n```\n\n### illustration-story\n```\nUse case: illustration-story\nPrimary request: 4-panel comic about a pet left alone at home\nScene/backdrop: cozy living room across panels\nSubject: pet reacting to the owner leaving, then relaxing, then returning to a composed pose\nStyle/medium: comic illustration with clear panels\nComposition/framing: 4 equal-sized vertical panels, readable actions per panel\nConstraints: no text; no logos or trademarks; no watermark\n```\n\n### stylized-concept\n```\nUse case: stylized-concept\nPrimary request: cavernous hangar interior with tall support beams and drifting fog\nScene/backdrop: industrial hangar interior, deep scale, light haze\nSubject: compact shuttle parked near the center\nStyle/medium: cinematic concept art, industrial realism\nComposition/framing: wide-angle, low-angle\nLighting/mood: volumetric light rays cutting through fog\nConstraints: no logos or trademarks; no watermark\n```\n\n### historical-scene\n```\nUse case: historical-scene\nPrimary request: outdoor crowd scene in Bethel, New York on August 16, 1969\nScene/backdrop: open field with period-appropriate staging\nSubject: crowd in period-accurate clothing, authentic environment\nStyle/medium: photorealistic photo\nComposition/framing: wide shot, eye-level\nConstraints: period-accurate details; no modern objects; no logos or trademarks; no watermark\n```\n\n## Asset type templates (taxonomy-aligned)\n\n### Website assets template\n```\nUse case: \nAsset type: \nPrimary request: \nScene/backdrop: \nSubject:
\nStyle/medium: \nComposition/framing: \nLighting/mood: \nColor palette: \nConstraints: \n```\n\n### Website assets example: minimal hero background\n```\nUse case: stylized-concept\nAsset type: landing page hero background\nPrimary request: minimal abstract background with a soft gradient and subtle texture\nStyle/medium: matte illustration / soft-rendered abstract background\nComposition/framing: wide composition with usable negative space for page copy\nLighting/mood: gentle studio glow\nColor palette: restrained neutral palette\nConstraints: no text; no logos; no watermark\n```\n\n### Website assets example: feature section illustration\n```\nUse case: stylized-concept\nAsset type: feature section illustration\nPrimary request: simple abstract shapes suggesting connection and flow\nScene/backdrop: subtle light-gray backdrop with faint texture\nStyle/medium: flat illustration; soft shadows; restrained contrast\nComposition/framing: centered cluster; open margins for UI\nColor palette: muted neutral palette\nConstraints: no text; no logos; no watermark\n```\n\n### Website assets example: blog header image\n```\nUse case: photorealistic-natural\nAsset type: blog header image\nPrimary request: overhead desk scene with notebook, pen, and coffee cup\nScene/backdrop: warm wooden tabletop\nStyle/medium: photorealistic photo\nComposition/framing: wide crop with clean room for page copy\nLighting/mood: soft morning light\nConstraints: no text; no logos; no watermark\n```\n\n### Game assets template\n```\nUse case: stylized-concept\nAsset type: \nPrimary request: \nScene/backdrop: (if applicable)\nSubject:
\nStyle/medium: ; \nComposition/framing: ; ; \nLighting/mood: