Data protection

Tool and data controls

Autessa AI Agent Firewall gives every agent tool allowlists, allowed sites for outbound data, approval for new email domains, blocking of secrets and holding of personal data leaving, a cap on rows read per run, and loop detection.

Turn on the firewall

Agent · Support Assistant · Tool and data checksSAMPLE

send_email → partner-new.example.ionew email domainREQUIRE_APPROVAL

http_post → hooks.unknown.dev with API keysecret leavingBLOCK

export_customers called 5× with same requestloopBLOCK

query_orders read 12,400 rowsrows-read cap 5,000BLOCK

lookup_order → api.shop.example.comallowed siteALLOW

The problem

Agents with tools are software that writes its own requests. Without guardrails at the tool boundary, one confused or hijacked run can email the wrong domain, post a secret to a webhook or read a whole table.

How tool and data controls works

STEP 1

Allowed tools

List the tools an agent may call. Anything else is blocked.

STEP 2

Allowed sites and email domains

List the sites tools may send data to; a new site needs approval. An email to a domain the agent hasn’t used before needs approval.

STEP 3

Secrets and personal data

A tool request that would send a secret to another site is blocked; personal data to another site needs approval; a reply containing a secret is blocked.

STEP 4

Rows and loops

Cap how many rows one run can read. The same tool called 5 times with the same request in one run is stopped.

Every setting

Allowed tools

Per agent allowlist; everything else blocked.

Allowed sites

Egress allowlist; new hosts need approval.

New email domains

Need approval the first time.

Secrets

Blocked from leaving in tool requests and replies.

Personal data (SSN, card numbers)

Sent to another site needs approval.

Rows read

Cap per agent run.

Loops

Five identical calls in one run are stopped.

Who can do what

Admins

Configure every control in the company or per-agent policy.

What to know

A secret about to leave in a tool request is a protected check that exceptions can never remove.
The same controls are available from the CLI: autessa agent-firewall policy ENFORCE --egress-allowlist a,b.

Tool and data controls: questions

Can I restrict which tools an AI agent can use?

Yes. List allowed tools per agent; anything else is blocked.

Can Autessa stop secrets leaking through an agent?

Yes. A tool request that would send a secret to another site is blocked, and so is a reply that contains one.

Turn on tool and data controls.

Allowlist the tools an agent may call and the sites it may send data to, hold new email domains for approval, block secrets leaving, cap rows read and stop tool loops. Start in Monitor and see what it would catch on your real traffic.

Turn on the firewall

Starts in Monitor · nothing changes for your customers until you enforce

100% customizable, code and all

Make Autessa AI Agent Firewall fully your own.

Use it out of the box, then change anything. Ask Autobot, Autessa’s AI builder, for a new field, screen, report, workflow or agent and it builds it into your copy of the app. Or open the code and edit it yourself. Nothing is locked: the app, its data and its code are yours.

Start free and make it yours