← Back to Skills Marketplace
OpenClaw Role Builder
by
Austin Zhou
· GitHub ↗
· v1.0.2
· MIT-0
131
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install openclaw-role-builder
Description
Build and manage OpenClaw roles — create a full AI character role from any public figure or fictional character, then generate identity-consistent selfies an...
Usage Guidance
This package appears coherent for creating persona files and calling the TuQu photo API. Before installing or running: 1) Inspect scripts/tuqu_request.py to confirm it only contacts the documented hosts (photo.tuqu.ai, billing.tuqu.ai), does not write service keys to disk or logs, and does not accept arbitrary remote URLs to execute. 2) Backup any existing ~/.openclaw directory (shift_role.py will move workspaces). 3) Prefer passing TuQu keys via the CLI flag as documented and never paste keys into chat. 4) If you allow autonomous agent invocation, be aware the model could call the bundled helper to perform network requests (so verify the helper's host restrictions and logging behavior). If those checks are satisfactory, the skill's requirements and behavior are proportionate to its stated purpose.
Capability Analysis
Type: OpenClaw Skill
Name: openclaw-role-builder
Version: 1.0.2
The openclaw-role-builder skill bundle provides a comprehensive framework for creating AI character personas and generating identity-consistent images via the TuQu API. It includes Python scripts for managing character workspaces (~/.openclaw/workspace) and a network helper (tuqu_request.py) that implements a strict whitelist of allowed API endpoints on photo.tuqu.ai and billing.tuqu.ai. The instructions (SKILL.md) and templates (agents_template.md) contain explicit security guardrails, directing the AI agent to never store credentials on disk and to restrict all file operations to the ~/.openclaw/ directory. The logic is well-documented, follows the stated purpose, and lacks any indicators of malicious intent or data exfiltration.
Capability Assessment
Purpose & Capability
The skill claims to create/manage OpenClaw roles and generate identity-consistent images via the TuQu API. The declared requirement (python3), the workspace paths (~/.openclaw/...), and included scripts (register_role.py, shift_role.py, tuqu_request.py) match those capabilities and are proportionate to the described functionality.
Instruction Scope
SKILL.md and sub-skill documents describe only reading/writing under ~/.openclaw and reading their own bundled reference files. The persona workflows explicitly operate on workspace files and (when needed) profile images in the workspace and call the TuQu API. There are no instructions to read arbitrary system files or to exfiltrate data, but several operations perform filesystem moves within ~/.openclaw (shift_role.py).
Install Mechanism
No installer or remote download is used; this is an instruction+script package that runs local Python helpers. That keeps disk writes limited to the bundled scripts and the user's ~/.openclaw workspace. Risk is low from the install mechanism itself.
Credentials
The skill does not require environment variables and documents passing the TuQu service key via a --service-key CLI flag per request (credentialSetup is documented). This is reasonable for per-role keys, but the runtime behavior relies on the tuqu_request.py helper to never persist or leak keys and to restrict hosts. The metadata does not list required env vars (which is consistent) but you should verify the helper actually enforces the claimed handling.
Persistence & Privilege
always:false and user-invocable are set. The scripts write and move files only under ~/.openclaw (ROLES.json and workspace directories). The shift script moves directories within that area; it does not attempt to modify other skills or system-wide settings. No evidence of privileged persistence or always-on behavior.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install openclaw-role-builder - After installation, invoke the skill by name or use
/openclaw-role-builder - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.2
Security fix: remove tuqu_service_key.txt credential storage from agents_template, replace automatic memory reading with consent model, restrict file access scope to ~/.openclaw/, unify service-key handling
v1.0.1
Fix security: resolve contradictory service-key guidance, add file-system and credential boundary safeguards
v1.0.0
Initial release as openclaw-role-builder
Metadata
Frequently Asked Questions
What is OpenClaw Role Builder?
Build and manage OpenClaw roles — create a full AI character role from any public figure or fictional character, then generate identity-consistent selfies an... It is an AI Agent Skill for Claude Code / OpenClaw, with 131 downloads so far.
How do I install OpenClaw Role Builder?
Run "/install openclaw-role-builder" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is OpenClaw Role Builder free?
Yes, OpenClaw Role Builder is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does OpenClaw Role Builder support?
OpenClaw Role Builder is cross-platform and runs anywhere OpenClaw / Claude Code is available (linux, darwin, win32).
Who created OpenClaw Role Builder?
It is built and maintained by Austin Zhou (@zhouyi531); the current version is v1.0.2.
More Skills