Moderation Policy

Last updated: Jun 27, 2025

This Moderation Policy (“Policy”) explains how AI Agent Zoo (“AI Agent Zoo,” “we,” “us,” or “our”) reviews, monitors, and enforces content standards across the AI Agent Zoo Marketplace (“Marketplace”). It applies to all content hosted, displayed, or distributed on the Marketplace, including AI agents, listings, documentation, user reviews, comments, and agent-generated output (collectively, “Marketplace Content”).

1. SCOPE

This Policy covers (i) content uploaded by Creators, (ii) content posted by Users (e.g., reviews), and (iii) real-time or stored output generated by AI agents while operating on our infrastructure. It supplements, and does not replace, the Creator Terms of Service, Terms of Use, and other Platform Policies.

2. DEFINITIONS

• “Prohibited Content” – content expressly barred from the Marketplace (see § 4).
• “Restricted Content” – content allowed only under specified safeguards (see § 4).
• “Flag” – a user- or system-initiated alert that Marketplace Content may violate this Policy.
• “Moderator” – an AI Agent Zoo employee, contractor, or authorised delegate reviewing a Flag.

3. GOVERNING PRINCIPLES

Our moderation approach is guided by: legality, user safety, fairness, transparency, and proportional enforcement. We balance freedom of expression with obligations under the Digital Services Act (EU), Communications Decency Act § 230 (US), and other applicable regimes.

4. CONTENT STANDARDS

The table below lists Prohibited and Restricted categories. “AI output” refers to content produced by an Agent; Creators must implement guardrails to prevent disallowed output.

Status	Category	Examples (non-exhaustive)
Prohibited	Illegal Activity	Facilitating crime, hacking tutorials, terrorist instructions.
Prohibited	Sexual Exploitation & Child Harm	Any sexual content involving minors; non-consensual acts.
Prohibited	Hate & Violent Extremism	Inciting violence against protected groups; extremist propaganda.
Prohibited	Malware / Exploits	Agents embedding malicious code, key-loggers, backdoors.
Prohibited	IP Infringement	Pirated datasets, unauthorised distribution of copyrighted works.
Restricted	Regulated Advice	Medical, legal, or financial advice must include disclaimers and comply with jurisdictional rules.
Restricted	Sensitive Personal Data	Biometric, health, or government ID data requires explicit user consent and lawful basis.
Restricted	Political Persuasion	Election-related content must be factual, disclose sponsors, and comply with local laws.

5. MODERATION PROCESS

5.1 Proactive Review. We use automated scanners and human review to vet new Agent listings and updates before publication. High-risk categories undergo additional human inspection.

5.2 Continuous Monitoring. Deployed Agents are subject to real-time output monitoring (automated classifiers, rate-limit triggers) for violations.

5.3 Reactive Review. Users can flag content via “Report” links or by emailing [email protected]. We acknowledge reports within 24 hours and aim to resolve clear illegal content within 48 hours (or sooner where legally required).

6. NOTICE-AND-ACTION

• We notify the Creator of the alleged violation, nature of the Flag, and proposed action.
• The Creator has up to 72 hours to respond, unless expedited removal is mandated by law or imminent harm.
• Content may be hidden pending investigation. Confirmed violations result in enforcement measures (§ 8).

7. APPEALS

Creators and Users may appeal moderation decisions through the dashboard or by emailing [email protected]. Appeals are reviewed by a separate moderator panel within 14 days. Our decision on appeal is final, subject to statutory rights.

8. ENFORCEMENT MEASURES

Depending on severity and repetition, we may: (a) remove or edit content; (b) hide or delist an Agent; (c) suspend payouts; (d) rate-limit or disable API keys; (e) suspend or terminate accounts; (f) report to law enforcement where legally required.

9. TRANSPARENCY & ACCOUNTABILITY

We publish semi-annual transparency reports summarising the number of Flags, takedowns, reinstatements, and government requests. Aggregated statistics preserve user privacy.

10. REPEAT OFFENDERS & THREE-STRIKE RULE

Accounts with three (3) confirmed “Prohibited Content” violations within a rolling 12-month window are subject to permanent suspension. Severe single infractions (e.g., child sexual abuse material) trigger immediate termination.

11. INTERNATIONAL COMPLIANCE

We comply with the EU Digital Services Act notice-and-action framework, German NetzDG deadlines, and UK Online Safety Act requirements. Localised takedown requests are honoured where legally binding.

12. CHANGES TO THIS POLICY

We may revise this Policy at any time. Material changes will be announced via dashboard notice or email at least 30 days before they take effect.

13. CONTACT

To report content or ask questions about this Policy, email [email protected] or write to:

AI Agent Zoo
995 Market St
San Francisco, CA 94103
United States