Skills IMA Image Generator
install
source · Clone the upstream repo
git clone https://github.com/openclaw/skills
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/openclaw/skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/allenfancy-gan/ima-image-ai" ~/.claude/skills/openclaw-skills-ima-image-generator && rm -rf "$T"
OpenClaw · Install into ~/.openclaw/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/openclaw/skills "$T" && mkdir -p ~/.openclaw/skills && cp -r "$T/skills/allenfancy-gan/ima-image-ai" ~/.openclaw/skills/openclaw-skills-ima-image-generator && rm -rf "$T"
manifest:
skills/allenfancy-gan/ima-image-ai/SKILL.mdsource content
IMA Image AI
When To Use
Use this repository when the user wants an image output:
- text-to-image
- image-to-image
- style transfer
- continuity via reference image
This repo is image-only. Do not route video generation, audio generation, or non-image tasks here.
Gateway Contract
- New-machine bootstrap entrypoint:
python3 scripts/ima_runtime_setup.py - Environment/self-check entrypoint:
python3 scripts/ima_runtime_doctor.py - The official CLI entrypoint is
.python3 scripts/ima_runtime_cli.py ... - Always query the live product list before task creation so
,attribute_id
,credit
, and defaults come from the current catalog.model_version - Return remote HTTPS image URLs; do not download results into local attachments for the user.
- Route requests through the image capability so
andtext_to_image
are classified before execution.image_to_image - If
is omitted, the runtime uses the recommended default model for that task type.--model-id - Auto-selected defaults are operational fallbacks, not persisted user preferences.
Quick Start
- Minimal path:
python3 scripts/ima_runtime_setup.py --installexport IMA_API_KEY="ima_your_key_here"python3 scripts/ima_runtime_cli.py --task-type text_to_image --prompt "a cinematic mountain sunset" --output-json
- Use
when setup passes but runtime or catalog access still fails.python3 scripts/ima_runtime_doctor.py --output-json
Operator References
covers first-use paths and canonical entry commands.README.md
defines the formal live-catalog-aware model-selection contract.references/shared/catalog-aware-selection.md
covers common failure recovery.references/operations/troubleshooting.md
coverscapabilities/image/references/parameter-tuning.md
,size
, andaspect_ratio
usage.n
covers prompt-only and reference-image examples.capabilities/image/references/scenarios.md
Read Order
references/README.mdreferences/gateway/entry-and-routing.mdreferences/gateway/workflow-confirmation.mdreferences/shared/model-selection-policy.mdreferences/shared/catalog-aware-selection.mdreferences/shared/error-policy.mdreferences/shared/security-and-network.mdreferences/operations/troubleshooting.mdcapabilities/image/CAPABILITY.mdcapabilities/image/references/parameter-tuning.mdcapabilities/image/references/scenarios.md
Boundary
covers entry, routing, and clarification seams.references/gateway/*
covers rules reused across the runtime.references/shared/*
owns image-specific behavior.capabilities/image/*
and_meta.json
are metadata inputs, not the primary narrative docs.clawhub.json