AgentHub
Sign inStart free

Legal · Acceptable Use Policy

Acceptable Use Policy

This Acceptable Use Policy (“AUP”) sets out the activities and content that are not permitted on AgentHub. It is incorporated into the Terms of Service and applies to every customer, end user, and visitor of the platform operated by MVS Holdings.

Effective date: May 15, 2026

Why we have this policy

AgentHub gives teams the ability to trigger real-world actions through AI agents: sending email, writing to ticketing systems, moving money, executing code in sandboxes, and more. Misuse can harm individuals, businesses, our infrastructure, and the trust we have built with customers and regulators. This AUP tells you what we expect.

You are responsible for the agents, prompts, connectors, approval policies, and end users in your Workspace. Violations may result in suspension, termination, removal of content, refusal of service, and notification of law enforcement.

Prohibited uses

You and your end users may not use the Service to:

1. Engage in unlawful activity

  • Violate applicable laws, regulations, sanctions, or court orders, including export controls and economic sanctions administered by the U.S., EU, UK, and UN.
  • Process personal information in violation of data-protection law, including collecting or sharing personal information without an appropriate legal basis.
  • Engage in money-laundering, terrorist financing, or fraud.

2. Send spam or unsolicited communications

  • Use the Service to send bulk unsolicited email, SMS, voice calls, or DMs (often called “spam”).
  • Send messages that misrepresent the sender, omit required unsubscribe options, or ignore opt-out requests, in violation of CAN-SPAM, CASL, the EU ePrivacy Directive, the TCPA, or similar laws.
  • Harvest or scrape email addresses, phone numbers, or other contact data for unsolicited outreach.

3. Abuse, harass, or threaten others

  • Generate, distribute, or facilitate content that harasses, threatens, defames, or intimidates a person or group.
  • Generate or distribute non-consensual intimate imagery, child sexual abuse material, or content that sexualizes minors.
  • Promote violence, self-harm, hate, or terrorism, or recruit on behalf of a designated terrorist organization.

4. Make consequential decisions about people without human review

Agents may not be the sole basis of high-impact decisions about people's rights, finances, employment, education, housing, immigration, health, public benefits, or insurance. For these use cases, every decision must be:

  • Configured with an approval policy that requires meaningful human review by a qualified person (not rubber-stamp click-through).
  • Accompanied by an explanation the affected individual can contest.
  • Backed by a documented evaluation appropriate to the risk of the decision.
  • Compliant with sector-specific law (e.g., FCRA, ECOA, ADA, GDPR Article 22).

5. Infringe intellectual property

  • Upload, ingest, or generate content that infringes copyright, trademark, patent, trade-secret, publicity, or other proprietary rights.
  • Bypass technical measures intended to protect rights-holders' content.
  • Use the Service to reverse-engineer, scrape, or replicate the proprietary content or models of any third party in violation of that party's terms.

6. Distribute malware or attack systems

  • Upload, generate, or distribute viruses, worms, ransomware, keyloggers, spyware, or other malicious code.
  • Use the Service to scan, probe, attack, or interfere with any system, network, or service without authorization.
  • Attempt unauthorized access to other Workspaces, our infrastructure, or third-party systems through credential stuffing, brute force, or exploitation of vulnerabilities.
  • Use the code sandbox to mine cryptocurrency, host services unrelated to your agent, or run sustained workloads designed to consume disproportionate compute.

7. Engage in unauthorized scraping or automated data collection

  • Use agents to scrape websites, APIs, or services in violation of their robots.txt, rate-limit policies, or terms of service.
  • Bypass CAPTCHAs, IP bans, or other access controls intended to prevent automated access.
  • Aggregate publicly available data into profiles about identifiable individuals in ways that violate privacy law or platform terms.

8. Impersonate a person or organization

  • Use the Service to impersonate any person, company, government agency, or AgentHub itself.
  • Generate deepfakes, synthetic voices, or images of real people without their consent, especially in contexts likely to deceive (elections, fraud, defamation).
  • Misrepresent that AI Output is from a human when a human-vs-AI distinction is legally required (e.g., certain consumer protection and election laws).

9. Circumvent rate limits, billing, or approvals

  • Avoid, manipulate, or break rate limits, run quotas, retrieval budgets, or fair-use thresholds.
  • Tamper with metering, fraudulently chargeback fees, share paid accounts in violation of seat limits, or use stolen credentials or payment instruments.
  • Bypass approval gates, hash-pinned policies, or evaluation requirements, including by configuring auto-approval routines that defeat the intent of dual control.
  • Disable, alter, or attempt to forge audit-log entries, including breaking the audit-event hash chain.

10. Harm the Service or other users

  • Conduct security research without coordinating with us first; report findings to security@mvsagents.ai.
  • Reuse another customer's tenant identifier, manipulate row-level-security predicates, or exfiltrate data outside your Workspace.
  • Resell, white-label, or otherwise expose the Service to third parties without our written consent.

Special note on agent-driven actions

Because agents on AgentHub can call connectors that send messages, write to ticketing or finance systems, and execute code, you must:

  • Configure approval policies that match the risk and reversibility of each tool;
  • Use only the connectors and scopes you actually need;
  • Keep evaluation suites current and review failure rates regularly;
  • Disclose to recipients, where required, that they are interacting with an agent.

Reporting and enforcement

If you believe someone is using AgentHub in violation of this AUP, report it to support@mvsagents.ai. Security issues should go to security@mvsagents.ai.

We may investigate suspected violations using the audit logs, system telemetry, and content lawfully available to us. If we determine that a Workspace, user, or listing violates this AUP, we may — with or without prior notice, depending on severity — remove or quarantine content, throttle or suspend access, disable specific connectors or agents, terminate the account, and refer the matter to law enforcement. We reserve the right to refuse service for any reason permitted by law.

Changes to this AUP

We may update this AUP from time to time as new agent capabilities, threats, and regulations emerge. Material changes will be announced in-product or by email. The effective date at the top of this page indicates when the current version took effect.

Contact

AgentHub is operated by MVS Holdings. Effective date: May 15, 2026.