Build and manage OpenClaw roles — create AI character personas and generate identity-consistent photos — installable via ClawHub or git clone.
| Capability | Sub-skill | Example triggers |
|---|---|---|
| ------------ | ----------- | ----------------- |
| Character creation — build a full 6-file persona (identity, soul, voiceprint, …) for any public figure or fictional character | openclaw-character-creator | 创建角色, 创建clone, 新建人物, "create a character" |
| Character switching — swap the active character in your workspace | openclaw-character-creator | /shift, 切换角色 |
| Identity-consistent selfies & portraits — generate images where the character's face stays consistent | tuqu-photo-api | 自拍, 写真, 发张图, "take a selfie" |
| Freestyle image generation — landscapes, objects, edits without a character in frame | tuqu-photo-api | 风景, 物品, 角色出镜, edit-only |
| Prompt enhancement, presets, billing — discover styles, improve prompts, check balance, recharge | tuqu-photo-api | "enhance prompt", "check balance" |
clawhub install openclaw-role-builder
npx skills add tuquai/openclaw-role-builder --all
Or install individual sub-skills:
npx skills add tuquai/openclaw-role-builder --skill openclaw-character-creator
npx skills add tuquai/openclaw-role-builder --skill tuqu-photo-api
git clone https://github.com/tuquai/openclaw-role-builder.git ~/.openclaw/skills/openclaw-role-builder
| Requirement | Why |
|---|---|
| ------------- | ----- |
| python3 | Required by tuqu-photo-api for the tuqu_request.py helper |
| TuQu service key | Needed for authenticated image-generation calls — obtain from |
| OpenClaw | The host agent that loads and executes these skills |
openclaw skills list
# Both openclaw-character-creator and tuqu-photo-api should appear
Once installed, start a new OpenClaw session and try:
Build high-consistency AI character personas with a full 6-file model (3 auxiliary + 3 main). Supports both well-known public figures and fictional/original characters, with workspace scaffolding and global role registration.
See openclaw-character-creator/SKILL.md for full usage.
Generate identity-consistent selfies, portraits, group photos, and freestyle images for OpenClaw characters via the tuqu.ai API. Covers prompt enhancement, preset discovery, character management, billing, and recharge flows.
See tuqu-photo-api/SKILL.md for full usage.
This skill collection writes to the following paths:
| Skill | Path | What is stored |
|---|---|---|
| ------- | ------ | ---------------- |
| openclaw-character-creator | ~/.openclaw/ROLES.json | Character names and workspace paths (no credentials) |
| openclaw-character-creator | ~/.openclaw/workspace-*/ | Character persona files and workspace state |
| openclaw-character-creator | ~/.openclaw/workspace/ | Active character workspace (managed by /shift) |
No service keys, tokens, or other credentials are persisted to disk by any script in this collection.
TuQu service keys are only used at runtime via --service-key CLI flags and are never written to config files.
~/.openclaw/ (the paths listed above). Noskill in this collection writes outside that directory.
references/ directory) and character workspace files under ~/.openclaw/. No arbitrary
filesystem reads.
--service-key CLI flag to tuqu_request.py. The script does not read credentials from environment variables,
config files, or any persisted store. Keys must never be written to disk.
scripts/tuqu_request.py, which restricts requests to the known hosts (photo.tuqu.ai, billing.tuqu.ai). No ad-hoc curl, wget, or
other HTTP tools.
mkdir for workspacescaffolding are executed. No arbitrary shell commands.
The openclaw-character-creator skill expects tuqu-photo-api to be installed as well. Install both for full functionality.
共 1 个版本