ERNIE-Image文生图SKILL

Generate images with ERNIE-Image. Use for image create requests incl. edits. Supports text-to-image ; - 1024x1024/1376x768/1264x848/ 1200x896/896x1200/848x12...

Security Scan

The skill appears to implement an ERNIE-Image client and will send prompts to Baidu's API, but there are several inconsistencies between the documentation/registry metadata and the actual script (missing declared env vars, mislabeled CLI help, and small hygiene issues) that merit attention before installing.

Purpose & Capability

Name/description match the code: a Python wrapper that calls an OpenAI-compatible ERNIE-Image endpoint. However the skill metadata declares no required environment variables or primary credential, while the script and SKILL.md clearly expect an API key (ERNIE-Image_API_KEY / ERNIE_Image_API_KEY / BAIDU_API_KEY or --api-key). Also minor mismatches (arg help says 'Gemini API key', SKILL.md references different skill paths) indicate sloppy packaging but not obvious malicious intent.

Instruction Scope

SKILL.md and the script limit behavior to building a prompt, calling an external image-generation API, decoding base64, and writing a PNG to disk. There are no instructions to read unrelated system files, transmit arbitrary local files, or communicate with unexpected endpoints beyond the documented Baidu AI Studio endpoint. The skill explicitly warns about CLI exposure of keys and instructs where files are written.

Install Mechanism

No automated install spec is present (instruction-only + a Python script). The script requires the openai Python package (SKILL.md tells users to pip install openai). This is low-risk compared to remote download/installers. No obscure URLs or extracted archives are used.

Credentials

Requesting an API key to call ERNIE-Image is proportionate. However the registry metadata does not declare the required env vars or a primary credential while the code expects environment variables (multiple possible names). The script also allows passing the key on the command line (which it warns may leak via process listings). These inconsistencies reduce transparency and increase the chance of accidental credential exposure.

Persistence & Privilege

The skill is not marked always:true, has no install that modifies other skills or system configs, and doesn't request persistent elevated privileges. It only writes generated images to the user-specified path.

What to consider before installing

This skill is plausibly what it claims (a small Python client for Baidu ERNIE-Image) but has packaging/documentation inconsistencies you should address before installing. Things to check/do: 1) Confirm you have a valid Baidu ERNIE-Image API key and set it as a user-scoped environment variable (ERNIE-Image_API_KEY or ERNIE_Image_API_KEY or BAIDU_API_KEY) rather than passing via --api-key to avoid exposure in process listings. 2) Inspect the script (scripts/generate_image.py) yourself — it posts the prompt to the base_url https://aistudio.baidu.com/llm/lmapi/v3; verify that URL is the official endpoint you intend to use. 3) Be aware the registry metadata did not declare required env vars or a primary credential — treat that as a packaging omission. 4) Install the openai Python package in an isolated environment (venv) before running. 5) Note the minor mismatches (arg help says 'Gemini API key', SKILL.md path hints differ from DESIGN.md) — they are likely sloppy but review code to be comfortable. If you need higher assurance, request the publisher to update metadata to list required env vars and fix the documentation mismatches, or run the script in a sandboxed account/environment first.

ERNIE-ImageImage Generation & Editing

Generate new images or edit existing ones using Baidu's ERNIE-Image API.

Prerequisites

• Clawdbot installed and configured

• Need Install python openai sdk: pip instsall openai

API Key

The script checks for API key in this order:

• 1.API key from https://aistudio.baidu.com/account/accessToken
• 2.ERNIE-Image_API_KEY environment variable
• 3.--api-key argument (use if user provided key in chat)

Usage

Run the script using absolute path (do NOT cd to skill directory first):

Generate new image:

python ~/.codex/skills/ERNIE-Image/scripts/generate_image.py --prompt "your image description" --filename "output-name.png" [--resolution 1024*1024|1366*768] [--api-key KEY]

Important: Always run from the user's current working directory so images are saved where the user is working, not in the skill directory.

Resolution Options

ERNIE-Image API supports three resolutions (uppercase K required):

• 1024x1024
• 1376x768
• 1264x848
• 1200x896
• 896x1200
• 848x1264
• 768x1376

Map user requests to API parameters:

• No mention of resolution → 1024x1024
• "low resolution", 1024x1024

If neither is available, the script exits with an error message.

Preflight + Common Failures (fast fixes)

• Preflight:

  • test -n \"$ERNIE-Image_API_KEY\" (or pass --api-key)
  • If editing: test -f \"path/to/input.png\"

• Common failures:

  • Error: No API key provided. → set ERNIE-Image_API_KEY or pass --api-key
  • “quota/permission/403” style API errors → wrong key, no access, or quota exceeded; try a different key/account

Filename Generation

Generate filenames with the pattern: yyyy-mm-dd-hh-mm-ss-name.png

Format: {timestamp}-{descriptive-name}.png

• Timestamp: Current date/time in format yyyy-mm-dd-hh-mm-ss (24-hour format)
• Name: Descriptive lowercase text with hyphens
• Keep the descriptive part concise (1-5 words typically)
• Use context from user's prompt or conversation
• If unclear, use random identifier (e.g., x9k2, a7b3)

Examples:

• Prompt "A serene Japanese garden" → 2025-11-23-14-23-05-japanese-garden.png
• Prompt "sunset over mountains" → 2025-11-23-15-30-12-sunset-mountains.png
• Prompt "create an image of a robot" → 2025-11-23-16-45-33-robot.png
• Unclear context → 2025-11-23-17-12-48-x9k2.png

Prompt Handling

For generation: Pass user's image description as-is to --prompt. Only rework if clearly insufficient.

Preserve user's creative intent in both cases.

Prompt Templates (high hit-rate)

Use templates when the user is vague or when edits must be precise.

• Generation template:
  • “Create an image of: <subject>. Style: <style>. Composition: <camera/shot>. Lighting: <lighting>. Background: <background>. Color palette: <palette>. Avoid: <list>.”

Output

• Saves PNG to current directory (or specified path if filename includes directory)
• Script outputs the full path to the generated image
• Do not read the image back - just inform the user of the saved path

Examples

Generate new image:

python ~/.codex/skills/ERNIE-Image/scripts/generate_image.py --prompt "A serene Japanese garden with cherry blossoms" --filename "2025-11-23-14-23-05-japanese-garden.png" --resolution 1024*1024

References

• https://aistudio.baidu.com/blog/detail/794723628346373
• https://aistudio.baidu.com/ernieimage
• https://aistudio.baidu.com/modelsdetail/46030/intro