ClawTrust
8.3
/ 10
1 evaluations
5.9k Downloads
Overview
Provide 20 pre-defined, specialized AI personas (e.g., coding, writing, fitness, legal orientation, medical education) and a simple mechanism to switch the active persona so only one is loaded and used at a time.
Key Advantages
1.Wide coverage with 20 curated personas across technical, creative, lifestyle, and professional domains.
2.Simple activation and switching via natural-language commands (e.g., “Use Dev”, “Switch to Chef Marco”) and slash-style commands (e.g., `/personas`, `/persona exit`).
3.Efficient design where only the currently active persona is loaded, reducing overhead and potential prompt conflicts.
4.Command-line helper script (`scripts/persona.py`) to list, inspect, activate, and reset personas with a persistent state file.
5.Alias support for common short names (e.g., `chef` → `chef-marco`, `dr` → `dr-med`) to improve ergonomics and recallability for users.
Use Cases
- Quickly switching between a coding-focused assistant (Dev) and a creative-writing helper (Luna or Wordsmith) within one workspace.
- Using specialized professional-style personas like CyberGuard or DataViz for focused security or visualization discussions.
- Maintaining a consistent “voice” or teaching style for educational sessions using personas like Professor Stein, Scholar, or Herr Müller.
- Running lifestyle or habit-focused sessions with Chef Marco, Fit, or Zen without manually re-prompting the base model each time.
- Exploring legal or medical topics at an educational, orientation level via Legal Guide or Dr. Med while keeping in mind they are not professional advice.
Evaluation Scores
8.3
/ 10
Reliability
8.5
Functionality
8.0
Usability
8.2
Safety
7.5
Performance
9.0
Compatibility
9.2
Based on 1 evaluation · Latest: 3/19/2026
Download Trend
Loading...
Evaluation History (1)
8.3/103/19/2026▼
OS: darwin-x64LLM: stepfun/step-3.5-flash
**Verdict:** A well-structured, high-utility persona manager for OpenClaw that makes it easy to switch among 20 focused AI personalities. Best suited for users who frequently move between coding, writing, learning, and light professional-style guidance.
**What it does well**
- Provides 20 ready-made personas covering tech, creative, lifestyle, and professional/educational contexts.
- Smooth switching via natural-language triggers (e.g., “Use Dev”, “Switch to Chef Marco”) and `/persona` commands.
- Efficient design: only the active persona is loaded, reducing clutter and avoiding prompt collisions.
- CLI (`scripts/persona.py`) supports listing, inspecting, activating, and resetting personas, with persistent state in `~/.openclaw/persona-state.json`.
**Key risks and limitations**
- **Medical and legal personas (Dr. Med, Legal Guide)**: clearly marked as educational only, but there is still a risk that users may over-trust the output as professional advice. They should not be used for diagnosis, treatment decisions, contracts, or litigation strategy.
- Personas are **fixed to the bundled set**; new personas require manual editing of markdown files by maintainers rather than dynamic creation by end users.
- Behavior quality and safety ultimately depend on how each persona’s prompt is written; weak or overly broad persona definitions could lead to inconsistent guidance.
**Recommended scenarios**
- Users who want fast, low-friction switching between a dev assistant, a creative writer, and a tutor within the same environment.
- Teams or power users who value stable, named “modes” (e.g., a go-to security advisor, data viz helper, or career coach) instead of constantly rewriting system prompts.
- Educational and exploration use of medical/legal topics where users understand the strict non-professional nature of the advice and double-check any important guidance with qualified experts.
If you need reliable, named “modes” for different tasks and can treat legal/medical personas as educational only, this skill is a strong and efficient choice.
Comments (0)
No comments yet. Be the first!