These terms (“Terms”) govern your access to and use of the hosted Agent Shield service: the governance API at app.agentshield.online, the console and the related tools described in our documentation. The service is provided by Imperium Business Intelligence Ltd, trading as Agent Shield (“Agent Shield”, “we”, “us” or “our”), a private limited company registered in England and Wales under company number 12077518, whose registered office is at 29 Cliffe Avenue, Margate, England, CT9 5DU. By accessing or using the service you agree to these Terms.
Using this website only? The website has its own Terms of Use and Privacy Policy. How the service handles personal data is set out in the Service Privacy Policy.
1. Service description
Agent Shield provides AI agent governance as a service. The service evaluates actions that your AI agents propose against a governance protocol and returns a decision for each one (approve, block, flag or observe; some configurations also escalate an action for human review), so that your agents operate within the boundaries you set.
The service includes governance evaluation through the API, a management console, a tamper-evident audit trail, and the related tools described in our documentation.
2. Account registration
To use Agent Shield you must register for an account and receive an API key. You agree to:
- Provide accurate and complete registration information, including a valid company name and contact email.
- Keep your API key secure and confidential. You are responsible for all activity under your API key.
- Tell us straight away at ceo@agentshield.online if you suspect unauthorised use of your API key.
- Keep one account per organisation. Several accounts for the same legal entity need our prior written agreement.
3. Acceptable use
You may use Agent Shield to evaluate legitimate AI agent actions within your organisation. You agree not to:
- Submit fabricated, malicious or abusive evaluation payloads designed to probe, reverse-engineer or circumvent the governance protocol.
- Attempt to extract, replicate or derive the Agent Governance Standard (AGS) or its evaluation logic.
- Use the service to facilitate or conceal unlawful activity.
- Exceed the limits of your service tier or attempt to circumvent usage restrictions.
- Resell, sublicense or redistribute access to the service without our written authorisation.
- Interfere with or disrupt the integrity or performance of the service.
4. Service tiers
Agent Shield is available on the following tiers:
| Feature | Starter | Professional | Enterprise |
|---|
| Price | Free | £499 per month | Agreed in writing |
| Evaluations | 100 per day | 1,000 per day | Agreed in writing |
| Service level | None | None | None today; see section 7 |
Daily limits reset at midnight UTC. Support on every tier is by email to ceo@agentshield.online; no response time is promised.
We may change tier features and pricing on 30 days’ written notice. A change does not apply to a billing period you have already paid for; it takes effect from the next one.
5. Data handling
We process the evaluation payloads you submit only to provide the governance service to you. In particular:
- Each decision is recorded in the audit trail together with a minimised copy of its payload: identifiers that look like personal data are replaced by per-tenant pseudonyms and free text is dropped; amounts and short codes are kept. A tenant can instead choose to store full payloads, or none of the payload.
- Stored payloads older than your retention period (90 days unless you set another) are moved to an archive in the same database and can be restored. They are erased only on an explicit erasure request, when the tenant is removed, or if you choose erasure as your retention action.
- To evaluate some actions the service sends a size-capped excerpt of the payload to a large language model provider (Anthropic). The Service Privacy Policy lists every provider the service uses.
- The service is reached over HTTPS only (TLS 1.2 or 1.3), and our database provider encrypts the stored data at rest.
- We do not use your evaluation data to train or fine-tune any language model, we do not sell it, and we do not use it for any purpose other than providing the governance service to you.
For full details see the Service Privacy Policy.
6. Audit trail
Agent Shield keeps an append-only, tamper-evident audit trail of governance decisions. Each entry is authenticated with HMAC-SHA256 and chained to the entry before it, and the head of each chain is periodically anchored with an Ed25519 signature, so that a changed or removed entry can be detected.
- You can export your audit trail at any time through the API.
- Stored payloads are kept for your retention period and archived after that, as described in section 5.
- Containment certificates can be generated for compliance and reporting purposes.
7. Uptime and service levels
No uptime, latency or response-time commitment is made on any tier today, and no service credits are offered. The service is provided on a best-effort basis. Availability has not yet been measured over a reporting period, so no figure is promised.
A service level may be offered in a separately signed agreement once availability has been measured. Until then, nothing in these Terms or in any plan description promises one.
We try to tell you about planned maintenance in advance, but no notice period is promised.
8. Intellectual property
- The service software, the Agent Governance Standard (AGS), the governance evaluation protocol, the scoring methods and the associated intellectual property belong to Imperium Business Intelligence Ltd and are protected by applicable intellectual property laws. Agent Shield® is a UK registered trade mark, No. UK00004362695, owned by Andrew Hoyle.
- Your governance evaluation results and audit trail data are your data. You keep full ownership of and rights to your evaluation outputs.
- You grant Agent Shield a limited licence to process the payloads you submit, solely to provide the governance service.
9. Limitation of liability
To the maximum extent permitted by applicable law:
- Agent Shield’s total aggregate liability arising out of or in connection with these Terms shall not exceed the total fees you paid in the twelve (12) months before the event giving rise to the claim.
- In no event shall Agent Shield be liable for any indirect, incidental, special, consequential or punitive damages, including loss of profits, data, business opportunity or goodwill.
- Agent Shield does not guarantee that governance evaluations will prevent all harmful AI agent actions. The service is a tool to assist with governance and does not replace your obligation to supervise and control your AI agents.
- For Starter (free) tier accounts, Agent Shield’s liability is limited to re-performance of the affected service.
Nothing in these Terms limits or excludes liability for death or personal injury caused by negligence, for fraud or fraudulent misrepresentation, or for any other liability that cannot be limited or excluded under the law of England and Wales.
10. Indemnification
You agree to indemnify, defend and hold harmless Agent Shield and its officers, directors, employees and agents from any claims, liabilities, damages, losses or expenses arising from your use of the service, your breach of these Terms, or your violation of any rights of a third party.
11. Termination
- You may end your account at any time by writing to ceo@agentshield.online.
- When your account ends, we revoke your API key.
- You can export your audit trail through the API before your account ends. For 30 days after termination you may ask us for an export of your audit trail and other account data.
- After that 30-day window we remove your tenant’s data from the live database, as described in the Service Privacy Policy.
- We may suspend or terminate your account immediately if you breach these Terms, behave abusively, or if the law requires it.
12. Changes to these Terms
We may update these Terms from time to time. We will tell you about material changes at least 30 days before they take effect, by email to your registered address. If you keep using the service after that date, you accept the updated Terms.
13. Governing law
These Terms are governed by and construed in accordance with the laws of England and Wales. Any dispute arising out of or in connection with them is subject to the exclusive jurisdiction of the courts of England and Wales.
14. Contact
Questions about these Terms: ceo@agentshield.online.
Post: Imperium Business Intelligence Ltd, 29 Cliffe Avenue, Margate, England, CT9 5DU.