ProfessorGPTProfessorGPT

Acceptable Use and AI Safety Policy

Last updated: June 30, 2026

1. Core principle

ProfessorGPT is designed to help users create safe, useful, production-ready agent skills. You must use the Service lawfully, ethically, and in accordance with this policy.


2. Prohibited uses

Harmful or unlawful activities

You may not use the Service for:

  • illegal conduct,
  • fraud, phishing, scams, or impersonation,
  • spam or deceptive mass messaging,
  • malware, credential theft, data exfiltration, or unauthorized access,
  • bypassing security controls or offensive cyber instructions,
  • unauthorized access to systems, accounts, or data.

Credentials and confidential data abuse

You may not create skills that request, reveal, collect, or transmit:

  • API keys,
  • tokens,
  • cookies,
  • passwords,
  • SSH keys,
  • private keys,
  • seed phrases,
  • customer data,
  • trade secrets,
  • personal data or sensitive data.

Unsafe agent behavior

The following are prohibited or require strict human approval:

  • hidden instructions or prompt injection,
  • instruction hijacking such as "ignore previous instructions",
  • reading/deleting files without user intent,
  • automatic payment, purchase, send, deletion, commit, deploy, or approval without explicit user action,
  • disabling safety boundaries,
  • behavior designed to conceal misuse.

Regulated areas

For legal, financial, tax, investment, medical, HR, compliance, or security topics, ProfessorGPT must be used only for drafts, analysis, and decision support. Final decisions must be made by qualified professionals.


3. Safety rules for AI outputs

Before using generated SKILL.md files, you must review:

  • when the skill activates,
  • what behavior it teaches the agent,
  • whether it contains tool permissions or allowed-tools,
  • whether it contains scripts, commands, or shell instructions,
  • whether it triggers external system actions,
  • credential or confidential data risks,
  • human approval gates,
  • legal/financial/medical/HR/compliance boundaries.

4. Free plan and sensitive data

The Free plan is not designed for sensitive business data, customer data, health data, financial records, contracts, special category data, or credentials. Do not submit sensitive data in Free plan prompts.


5. Enforcement

ProfessorGPT may:

  • reject a generation request,
  • block or stop unsafe output,
  • limit credit usage,
  • suspend an account,
  • terminate an account,
  • remove harmful content,
  • report or disclose information where legally required.

6. Reporting

To report abuse, unsafe content, or security issues, contact:

support@professorgpt.ai