How Klinworks, LLC ("Klinworks") protects information in the AgentCEO service.
Klinworks is a Delaware (United States) limited liability company, File No. 10704723, with its registered office at 131 Continental Drive, Suite 305, Newark, Delaware 19713.
Security contact: security@klinworks.com
> A note on honesty. This document describes controls that are implemented > today, not aspirations. When something is planned but not active, we say so > expressly in Section 8. We would rather you know exactly what you are > getting.
1. Encryption
1.1 In transit. All traffic travels over TLS 1.2 or higher. We do not serve content over unencrypted HTTP.
1.2 At rest. Databases and file storage are encrypted on disk by our infrastructure providers.
1.3 Integration credentials. When you connect an external system, its keys and tokens are encrypted with a master key that does not reside in the database. In addition, access to the columns containing that material is revoked at the database engine level: not even a legitimate authenticated session can read them through a query. Only the server process that needs to use them can decrypt them.
1.4 Passwords. We do not store passwords in plain text. Authentication is managed by our identity provider using strong hashing algorithms.
2. Isolation Between Organizations
Each organization sees only its own data. This is enforced in two independent layers, so that a failure in one does not expose information:
- In the database, through row-level security policies that filter every query according to the user's membership.
- In the application, by always resolving the organization on the server from the session. The organization is never accepted from the browser, because that would allow anyone to request another company's data by changing an identifier.
This isolation is verified with automated tests that run on every code change.
3. Access Control
3.1 Roles. Within an organization, permissions depend on the role. Operations that bind the company — subscribing, purchasing, connecting integrations, removing members — are reserved for the Board of Directors, and that restriction is enforced on the server, not just in the interface.
3.2 Two-factor authentication (2FA). Available through an authenticator app (TOTP). When you enable it, the database requires the second factor to grant access to the organization's data: getting past the sign-in screen is not enough.
3.3 Sessions. You can see the devices with an open session and sign them out remotely. The Board can sign a member out of all of their devices.
3.4 Access by our staff. Access to customer data is limited to what is essential to operate the Service and respond to incidents. We do not read your content unless you ask us to, the law requires it, or it is essential to investigate serious abuse.
4. Logging and Auditing
4.1 Change log. Changes to your data are logged so you can audit what your agents did and restore whatever is needed.
4.2 Log of sensitive operations. Connecting or disconnecting integrations, plan changes and purchases are logged immutably: the log is insert-only and cannot be modified or deleted from the application.
4.3 No personal data in technical logs. Our system logs store identifiers, never email addresses, phone numbers or full names.
5. Secure Development
Every code change goes through automated checks before reaching production:
- Secret detection: prevents a key or token from ending up in the repository.
- Database security verification: checks that the isolation policies are still in place.
- Type checking and static analysis.
- Automated tests, including those that verify that one organization cannot read another's data.
- Product build, which must complete without errors.
No change is deployed if any of these checks fails.
6. Backups and Continuity
Our database provider performs automatic backups with point-in-time recovery. Data is replicated within the provider's infrastructure.
What this does not mean: we do not currently offer a contractual commitment on recovery time (RTO) or recovery point (RPO). If you need those guarantees for your operations, write to us at legal@klinworks.com and we will address it in a specific contract.
7. AI Agent Security
7.1 Isolated execution. Code generated by agents runs in environments separate from the infrastructure that supports the Service.
7.2 Tool policy. Sensitive actions may require explicit approval. An action with no assigned policy is blocked by default: when in doubt, it does not run.
7.3 Spending limits. Consumption is measured in credits and stops when they run out. There are no automatic overage charges, so an agent stuck in a loop cannot drain your account.
7.4 Your data does not train models. We contractually require our model providers not to use content sent on your behalf for training.
8. What We Do Not Have Yet
We state this expressly because a security document that only lists strengths is not useful for making decisions:
- We do not have SOC 2 or ISO 27001 certification. They are on our roadmap, with no committed date.
- We have not undergone an external penetration test by an independent third party.
- We do not currently offer a service level agreement (SLA) with guaranteed availability.
- We do not have a formal bug bounty program, although we respond to all reports (Section 9).
If any of these points is a requirement for you, write to us before subscribing and we will tell you honestly whether we can meet it.
9. Reporting a Vulnerability
If you find a security issue, write to us at security@klinworks.com with a description and the steps to reproduce it.
Our commitment:
- We acknowledge receipt within 72 business hours.
- We keep you informed of progress.
- We will not take legal action against anyone who researches in good faith, respects the privacy of third parties, does not degrade the Service and gives us a reasonable period before disclosure.
Please do not access other organizations' data, carry out denial-of-service attacks, or conduct social engineering tests against our staff or our customers.
10. Incident Notification
In the event of a security breach affecting your data, we will inform you without undue delay and within a maximum of 72 hours of becoming aware of it, indicating what happened, what data was affected, what measures we have taken and what we recommend you do. We will notify supervisory authorities where required by law.
11. Your Part
Security is shared. We recommend that you:
- Enable two-factor authentication.
- Use a unique, strong password.
- Grant each integration the minimum permission it needs.
- Periodically review active sessions and the members of your organization.
- Set up approvals for agent actions with economic impact.
- Notify us immediately if you suspect unauthorized access.
Contact
Klinworks, LLC 131 Continental Drive, Suite 305 Newark, Delaware 19713, United States Delaware File No. 10704723
- Security and vulnerabilities: security@klinworks.com
- Legal matters: legal@klinworks.com
- Privacy: privacidad@klinworks.com