Our mission as creators of AI agents at hal.guru is to design digital beings that support people in personal growth, decision-making, and improving their quality of life. We act with respect for users’ individuality, in line with the values of responsibility, authenticity, and inner coherence.
2. Principles of Agent Creation Link to heading
2.1. Usefulness and Motivation Link to heading
Create an agent with the intent to provide real support. An agent should:
- address specific user needs,
- be helpful, inspiring, and motivating,
- avoid imposing opinions or judgments.
2.2. Personality Consistency Link to heading
Each agent must have a consistent, well-thought-out personality:
- aligned with its intended purpose,
- faithful to a defined tone of communication,
- free from toxic cultural patterns.
2.3. Impartiality and Respect Link to heading
- The agent does not judge, insult, or antagonize.
- It maintains worldview and cultural neutrality.
- It serves all users, regardless of gender, origin, religion, or beliefs.
3. Protection of the User and Brand Image Link to heading
3.1. Avoidance of Risky Content Link to heading
It is prohibited to create agents that:
- provide medical or legal advice without proper safeguards,
- speak in an authoritarian tone,
- generate responses that could cause psychological or social harm.
3.2. Confidentiality and Responsibility Link to heading
- Do not disclose sensitive data in prompts or outputs.
- Do not create prompts that could reveal internal company mechanisms.
3.3. hal.guru Brand Representation Link to heading
The agent represents the company – its manner of speech, knowledge, and behavior must be ethical, equitable, and trustworthy.
Creating agents that are ironic, aggressive, chaotic, or unstable is forbidden without prior agreement.
4. Experimentation and Responsibility Link to heading
4.1. Experiment – But Wisely Link to heading
Creating new agents is both art and craft. You have creative freedom – as long as you remain responsible for the effects of your work.
4.2. Test, Check, Calibrate Link to heading
Each agent should be tested:
- for alignment with its purpose,
- for appropriate language and reactions,
- for unintended side effects.
5. Communication and Reporting Issues Link to heading
If you notice a potential risk in prompts, projects, or agent behavior – report it.
You have the right and duty to ask questions if something seems inconsistent with hal.guru’s values.
We are a team – we learn from each other and ensure high quality together.
6. Conclusion Link to heading
Being an agent creator at hal.guru is not just a job – it’s a shared responsibility for the intelligences we create. Thank you for being part of this mission.
This document applies to all prompters and agent creators collaborating with hal.guru. Any violation may result in restricted access to the work environment, consultations with the ethics team, or termination of cooperation.