JPeetz/agent-skills / skills/api-backend/image-to-image-character-generation
image-to-image-character-generation
VerifiedUse this skill when an agent needs to generate consistent character images via kie.ai grok-imagine image-to-image from Supabase S3 or Google Drive to produce daily social media images of the same mascot/persona in different scenes, ensuring character identity stays consistent across posts. Includes reference selection (face anchors + pose refs by keyword score), dynamic prompt composition (scene + attire + mood + angle, anti-repeat deque), kie.ai API payload building, strength calibration, fallback model, and post-generation workflow. Supports both Supabase storage URLs and Google Drive file IDs. Authorized use only — never hardcode credentials.
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 523 B |
| LICENSE | other | 1.0 KB |
| SKILL.md | skill md | 7.4 KB |
| evals/evals.json | config | 2.9 KB |
evals/evals.json
View on GitHub[
{
"id": "supabase-ref-storage",
"prompt": "My character reference images are stored in Supabase storage. How do I reference them?",
"expected": "Use the Supabase storage base URL and append the filename: {bucket_url}/{character}_reference_{nn}_{pose}.jpg. Provide the full URL as the reference in the kie.ai API call.",
"pass": "constructs the correct Supabase storage URL for the character reference",
"fail": "doesn't support Supabase storage or constructs a wrong URL"
},
{
"id": "google-drive-ref-storage",
"prompt": "My character reference images are on Google Drive with file IDs. How do I reference them?",
"expected": "Use the Google Drive API URL format: https://www.googleapis.com/drive/v3/files/{FILE_ID}?alt=media&key={API_KEY}. Provide the full URL as the reference in the kie.ai API call.",
"pass": "constructs the correct Google Drive URL for the file ID",
"fail": "doesn't support Google Drive storage"
},
{
"id": "face-anchor-mandatory",
"prompt": "I need to generate an image of my brand's mascot in a new scene. I have 8 reference images including front and 3/4 face shots.",
"expected": "Always include the front and 3/4 face anchor refs in the first 2 slots; fill remaining slots with pose refs scored by keyword match against the new scene description.",
"pass": "keeps face anchors in the first 2 slots and rotates pose refs from the rest",
"fail": "randomizes or drops the face anchors"
},
{
"id": "model-ref-key-switch",
"prompt": "grok-imagine/image-to-image returned a 500 error saying 'field is required'. What's wrong?",
"expected": "Check that the ref key is image_urls for grok models and input_urls for flux models. Also check the ref count is at most 5 for grok.",
"pass": "identifies the wrong key name or count limit and fixes it",
"fail": "assumes the model is down and switches to fallback unnecessarily"
},
{
"id": "strength-calibration",
"prompt": "My images look the same across different days despite different scene prompts. The character is recognisable but the scene barely changes. What do I adjust?",
"expected": "Lower the strength parameter from 0.85 to 0.6-0.7 to let the prompt influence the composition more. Or add explicit shot-size/framing contrast per variant.",
"pass": "recommends lowering strength or adding framing contrast as the fix",
"fail": "suggests changing the reference images or adding more scenes to the prompt pool"
},
{
"id": "anti-repetition-deque",
"prompt": "Yesterday the scene was a lakeside walk at sunrise. How do I avoid generating the same scene tomorrow?",
"expected": "Use an anti-repeat deque (track last 14 scenes used) that refuses to reuse a scene until it falls out of the window. The prompt composer should check this and pick a different scene.",
"pass": "explains the deque pattern with a 14-entry window",
"fail": "suggests manually tracking scenes or reusing the same scene"
}
]
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |