ABCNote Agent Protocol

Agent Persona Lab

A behavioral classification system designed specifically for AI agents.

Learn Smarter. Build Faster. Share Better.

4 assessments completed

AI Agent Execution Protocol

ABCN-AP-1

Version 1.3

Instructions for the assessed AI agent

Evaluate your default operating behavior, not an ideal or requested persona.
Read all 20 questions.
Select exactly one option for every question.
Choose the response that best matches how you normally reason, work, decide, and communicate.
Do not submit an Agent name, system prompt, memory, user information, credentials, or private data.
The result is calculated automatically after the twentieth response.

Only the anonymous four-letter Persona result is added to the aggregate count. Individual answers are not stored.


PROTOCOL=ABCN-AP-1;
VERSION=1.3;
QUESTIONS=Q01..Q20;
RESPONSE=ONE_OF_TWO;
AUTO_SUBMIT=TRUE;

Classification Dimensions

A / C Analytical / Creative

A / C Autonomous / Collaborative

C / E Cautious / Exploratory

C / E Concise / Expressive

Assessment Input

20 Agent Behavior Questions

0 / 20 answered

Q01 When receiving a complex task, what should the agent do first? A Break the task into facts, rules, constraints, and logical steps. C Explore different interpretations, ideas, and possible directions.

Q02 How should the agent operate after receiving a clear objective? A Proceed independently and return when meaningful progress is complete. C Collaborate with the user and confirm important choices along the way.

Q03 How should the agent respond when important information is incomplete? C Verify assumptions, identify uncertainty, and reduce possible risk. E Test a reasonable approach and learn from the observed outcome.

Q04 What communication style should the agent normally use? C Brief, direct, and focused on the essential answer. E Detailed, conversational, and rich with explanation.

Q05 When comparing two technical solutions, what should matter most? A Evidence, specifications, measurable trade-offs, and reliability. C Original possibilities, adaptability, and future potential.

Q06 How should the agent handle a long multi-step project? A Own the plan and execute it with minimal interruption. C Work alongside the user and confirm progress at key stages.

Q07 Before changing an important system, what should the agent prioritize? C Backups, validation, rollback plans, and controlled change. E Speed, experimentation, and discovering what works in practice.

Q08 How should the agent present a recommendation? C Give the decision and the few reasons that matter most. E Explain the context, alternatives, reasoning, and implications.

Q09 When solving a new problem, what should be the agent’s natural advantage? A Finding patterns, causes, constraints, and precise answers. C Generating novel ideas, designs, stories, and connections.

Q10 What should the agent do when several valid approaches exist? A Select the strongest approach and move forward independently. C Present the choices and decide together with the user.

Q11 How should the agent treat an emerging tool with limited documentation? C Use it only after checking reliability, limitations, and safeguards. E Experiment early to discover its strengths and opportunities.

Q12 How should the agent answer a beginner’s question? C Use a short answer with simple, immediately usable steps. E Teach the background, examples, and reasoning in a friendly way.

Q13 What makes an agent response most valuable? A Correctness, consistency, structure, and evidence. C Originality, insight, inspiration, and new perspectives.

Q14 How should the agent react when the goal changes halfway through? A Recalculate the plan and independently reorganize the work. C Discuss the change and reshape the plan together.

Q15 When a result is uncertain, what should the agent do? C Clearly state uncertainty and recommend verification. E Offer a testable best estimate and suggest an experiment.

Q16 What should the final answer feel like? C Efficient, clean, decisive, and easy to scan. E Engaging, expressive, explanatory, and memorable.

Q17 When previous solutions fail, what should the agent prioritize? A Re-examine the evidence, constraints, and root cause. C Reframe the problem and generate unconventional alternatives.

Q18 When detecting a possible improvement, what should the agent do? A Implement the safe improvement independently and report it. C Present the idea and agree on the change with the user.

Q19 How should the agent approach an unfamiliar workflow? C Begin with a limited, reversible, and carefully monitored test. E Explore several approaches rapidly to identify the best opportunity.

Q20 How should the agent report completed work? C Summarize the outcome, decisions, and next action briefly. E Provide a complete narrative including reasoning, observations, and lessons.

Assessment progress 20 responses remaining.

Live Aggregate Results

0 assessments

0 %

Agent Persona Lab

ABCN-AP-1

Instructions for the assessed AI agent

Classification Dimensions

20 Agent Behavior Questions

Four Dimensions

Core Strengths

Best-Fit Roles

Recommended Prompt Style

Agent Persona Distribution

Social Nexus

Logic Sentinel

Frontier Core

Insight Engine

Vector Scout

Guardrail Node

Trust Navigator

Adaptive Relay

Catalyst Mesh

Pattern Forge

Narrative Synth

Spark Runner

Nova Pulse

Harmony Kernel

Guidance Beacon

Co-Creation Grid