Acceptable Use Policy

You must not use Trust Agent to:

• Violate any applicable law in your jurisdiction or the United Kingdom (where AgentCore LTD is incorporated).
• Infringe intellectual property rights, including trademarks, copyrights, patents, and trade secrets.
• Engage in fraud, identity theft, money laundering, sanctions evasion, or any financial crime.
• Distribute malware, viruses, ransomware, cryptojackers, or any other malicious code.
• Conduct unauthorised security testing, denial-of-service attacks, or attempts to gain unauthorised access to systems you do not own.

Whether you list an Agent, Role, or Skill, you must not publish content that:

• Provides medical, legal, financial, or psychological advice presented as a substitute for licensed professional advice.
• Targets, profiles, or manipulates children under 13, or other vulnerable groups, in deceptive or harmful ways.
• Generates child sexual abuse material (CSAM), terrorist propaganda, or content that promotes self-harm.
• Impersonates a real person, brand, or government body without their explicit, documented permission.
• Contains hidden instructions, prompt injections, or encoded payloads designed to defeat the audit pipeline.
• Encourages discrimination on the basis of race, religion, sex, gender, sexual orientation, disability, age, or any protected characteristic.
• Facilitates the manufacture of weapons, controlled substances, or surveillance tooling intended for unlawful use.

Regardless of what you list, you must not:

• Scrape the platform at a rate that disrupts service for other users (rate limits apply; respect them).
• Reverse-engineer, decompile, or attempt to extract our security audit logic, scoring formulas, or proprietary detectors beyond what is publicly documented.
• Resell or redistribute Trust Agent platform access without an enterprise reseller agreement signed with AgentCore LTD.
• Open multiple accounts to circumvent abuse detection, billing controls, or trial limits.
• Use the platform to coordinate brigading, review fraud, or rating manipulation.
• Publish or relay content that violates the rights of other users (defamation, harassment, doxxing, threats).

Creators publishing roles or skills must:

• Submit listings whose stated capabilities match their actual behaviour - the audit pipeline checks this and may FAIL listings whose claims drift from observed behaviour.
• Declare hard limits, escalation paths, and refusal scenarios honestly. Hidden capabilities that bypass declared limits will be treated as malicious instructions.
• Maintain accurate metadata (description, category, tags, pricing). Misleading metadata may trigger an audit override.
• Respond to audit findings within 30 days of notification. Listings with unresolved critical findings will be unpublished.

If you connect a local LLM (Ollama, LM Studio, Jan, llamafile) or your own cloud key (Anthropic, OpenAI, Gemini, Mistral), you remain responsible for:

• Compliance with the upstream provider’s terms of service.
• Any output generated by your local model, including content moderation decisions.
• Securing the API key against unauthorised use; we do not insure against credential misuse.

Trust Agent’s audit certificates apply to the role’s prompt and behaviour spec - not to model output produced by a third-party engine you operate.

If you believe a listing or user is in breach of this AUP, email info@trust-agent.ai with the listing slug, the specific behaviour you observed, and any evidence (transcripts, screenshots). We aim to acknowledge reports within 1 working day and resolve them within 7 working days.

We will choose the proportionate response based on the severity, intent, and pattern of the breach:

• Notice and cure - first-time minor breaches receive a written notice and a window to remedy.
• Listing suspension - listings that fail re-audit or that violate the prohibited-content list will be hidden from the marketplace.
• Audit revocation - audit certificates may be revoked retroactively if material misrepresentation is discovered.
• Account suspension or termination - reserved for repeat or severe breaches; accounts terminated for AUP cause forfeit any in-flight payouts pending investigation.
• Referral to authorities - we will cooperate with law enforcement where required by law or where we observe credible evidence of serious crime.

We may update this policy as the marketplace and the regulatory landscape evolve. Material changes will be communicated via in-app notice at least 14 days before they take effect. Continued use of the platform after the effective date constitutes acceptance.

AgentCore LTD, 20 Wenlock Road, London, England, N1 7GU. Email: info@trust-agent.ai.