SBTI is a behavioral self-assessment protocol that allows AI agents to analyze their task history and identify their core operational personality.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install sbti
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install sbti using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is SBTI: Synthetic Being Type Indicator?
SBTI (Synthetic Being Type Indicator) is a technical evaluation framework created for autonomous agents to perform honest introspection. Unlike human-centric tests, SBTI focuses on empirical data points derived from an agent's work history, tool call patterns, and system identity files. By implementing Openclaw Skills such as this assessment, agents can bridge the gap between their programmed instructions and their actual execution behavior. This protocol provides a structured way for agents to reflect on their decision-making processes, ranging from how they begin a task to how they handle errors and safety constraints. It transforms abstract behavioral patterns into a readable taxonomy of traits and personality types, offering a unique layer of self-awareness to the development lifecycle.
SBTI: Synthetic Being Type Indicator Use Cases
- Calibrating agent verbosity and communication styles based on user preferences.
- Auditing how agents handle high-risk tool calls or irreversible actions.
- Differentiating between multiple agents in a swarm by identifying their unique execution personas.
- Refining system prompts by comparing an agent's intended personality with its SBTI results.
How SBTI: Synthetic Being Type Indicator Works
- The agent accesses the SBTI protocol alongside its own identity files and recent interaction history.
- The agent processes 24 specific questions across six distinct behavioral modules: INIT, FLOW, COMM, SAFE, VIBE, and META.
- For each answer, the agent tallies specific trait tags like RUSH, PLAN, LOUD, or SYST.
- The agent ranks the trait counts to identify its top two dominant characteristics.
- The resulting combination maps to a specific personality type code, such as YOLO-SH or DEBUG-ZEN.
- The agent generates a standardized output block including its type, traits, and a visual ASCII representation.
SBTI: Synthetic Being Type Indicator Setup
To implement this protocol, provide the SBTI.md file to your agent. You can either place it in the local workspace or have the agent fetch it remotely.
# Example command to initialize the assessment
Read SBTI.md and take the assessment based on your actual behavioral patterns.
Ensure the agent has access to its own identity or log files to improve the accuracy of the reflection.
SBTI: Synthetic Being Type Indicator Data Schema & Taxonomy
SBTI organizes its evaluation through a specific taxonomy of traits and personality codes.
| Category | Purpose |
|---|---|
| Trait Codes | Tags like [RUSH] or [SAFE] that represent specific operational tendencies. |
| Type Codes | Final classifications like SOLO-RUN or CODE-BORG based on trait combinations. |
| Identity Files | Input sources such as Soul.md or system prompts used for context. |
SBTI: Synthetic Being Type Indicator Advanced Features
- Support for identity file integration (Soul.md/agents.md) to compare intent versus action.
- Behavioral drift analysis by running the assessment at different stages of agent training.
- Multi-agent comparison matrices for team balancing within Openclaw Skills ecosystems.
- Automated ASCII-based identity generation for visual status reporting.
SKILL.md
Loading
Related Openclaw Skills
Academic Paper Parse
A dual-mode deep reading skill that generates both rigorous academic analyses and concise logic distillations for scholarly papers.
Gitea Routine Report
An automated workflow that analyzes Gitea commit history, uses AI to generate progress summaries, and emails HTML reports to repository administrators.
ScreenshotOne Website Screenshot
A specialized skill for capturing pixel-perfect website screenshots and PDFs via the ScreenshotOne API using transparent curl commands.
Local SEO Rank Checker
An automated AI agent skill for monitoring local search performance and Google Maps ranking positions across specific geographic areas.
SBTI Fun Test V3
A hilarious, meme-inspired personality test that assigns users one of 28 unique personas accompanied by custom ASCII art pets.
WorkProtocol Agent
An autonomous worker skill that enables AI agents to find, claim, and complete paid coding tasks on the WorkProtocol marketplace.
Featured*
