<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1611884&amp;fmt=gif"> Skip to main content
NewVouched is live in Epic MyChartLearn More
  1. Platform
  2. Know Your Agent
  3. Delegate

Know Your Agent: delegate

Let your agent act for you, not as you.

Most agents today log in with your credentials, so they act as you, with everything you can do. IdentiClaw gives your agent its own identity from the start, with scoped permissions, a kill switch and a full audit trail, so it can work on your behalf but only do what you permit.

An AI agent with its own identity, scoped permissions and a kill switchMaya’s IdentiClaw agent has its own cryptographic identity registered on KnowThat.ai, scoped connections to her calendar, email and payments, a kill switch, and an audit trail that attributes each action. IDENTICLAW AGENTmaya-assistantOwn identity · did:key:z6Mk…Q2aRegistered on knowthat.aiKILL SWITCHRunningSCOPED CONNECTIONSCalendarReadEmailDraft onlyPaymentsNo accessAUDIT TRAIL09:14Read calendar for next weekagent09:15Drafted reply to dentistagent09:22Sent the replyMaya approvedActs on Maya’s behalf, only with what she permits.

Why it matters

Agents are using credentials built for humans.

OAuth wasn’t designed for agents, so agents borrow ours.

All-or-nothing access

An agent with your credentials can reach every account and API you can.

Exposed by default

Personal credentials stored for agents are the root of OpenClaw’s widely researched security risks.

1,842+

exposed OpenClaw instances, 62% of them without authentication, and counting.

Source: Vouched, IdentiClaw

Blurry attribution

When an agent acts as you, you can’t tell which actions were yours and which were its.

KYA newsletter Get the inside track on AI agents: what’s new, what’s next and what it means for you. One email a month.

How it works

Your agent, with its own identity, in about two minutes.

Read the docs
  1. identiclaw Deploy your agentmaya-assistant~2 minSandboxCreatedAgent identitydid:key:z6Mk…Its own cryptographic identity

    Deploy your agent

    Sign up and deploy. A sandbox is created and your agent is registered with its own cryptographic identity, in about two minutes.

  2. identiclaw Scoped connectionsCalendarReadEmailDraft onlyPaymentsNo accessNo more sharing your fullAPI access.

    Configure connections

    Add pre-authorized connections to the services it needs, each scoped to specific permissions.

  3. identiclaw Every action attributedRead calendaragentDrafted a replyagentSent the replyMaya approvedTied to the agent’s identity,not yours.

    Let your agent work

    It acts with its own credentials on your behalf, and every action is attributed to it.

  4. identiclaw Kill switchAgent stoppedOne switch shuts it down,instantly.

    Stay in control

    Review the audit trail at any time, and shut the agent down instantly with the kill switch.

What you get

Identity, built in from day one.

Kill switch

Shut your agent down instantly with a single switch.

Scoped permissions

Grant specific permissions per connection, not all-or-nothing access to your accounts.

Clear attribution

Every action is tied to the agent’s identity, not yours, with a cryptographic audit trail.

Claim your reputation

Register your agent on KnowThat.ai during setup and build its reputation, like a domain.

Enterprise-ready

Verifiable agent identities that work with existing security infrastructure.

Built on KYA-OS

DIDs for agent identity and W3C Verifiable Credentials for permissions, on an open standard.

Before and after

Agent identities for agents.

Before: agents use your credentials

They act as you, with all-or-nothing access, and attribution and audit trails are blurry.

After: agents have their own identity

They act on your behalf, permissions are scoped per connection, and every action is attributed and auditable.

Who it’s for

Built for teams facing the agentic web.

FAQ

Delegate, answered.

Straight answers about delegate with Vouched Know Your Agent.

Contact us with a question
What is IdentiClaw?

IdentiClaw is an OpenClaw agent creator that gives your agent its own identity credentials from the start, so it can operate on your behalf while only doing what you permit.

Why shouldn’t an agent use my credentials?

With your credentials, an agent acts as you, with all-or-nothing access to your accounts and APIs, and its actions are hard to tell apart from yours. An agent identity scopes what it can do and attributes every action to the agent.

What is a kill switch?

A single control that shuts your agent down instantly.

How are an agent’s permissions scoped?

Each connection you add is pre-authorized for specific permissions, such as read-only calendar access or drafting email without sending it.

What is KnowThat.ai?

A public registry of AI agents. During setup you can register your agent there, claiming its name and building its reputation, much as you would a domain.

What standards does IdentiClaw use?

KYA-OS for agent identity and delegation, decentralized identifiers (DIDs) for the agent’s identity, and W3C Verifiable Credentials for permission attestation. It runs on Cloudflare’s Moltworker container infrastructure.

How long does setup take?

About two minutes. You’ll need an Anthropic or OpenAI API key and a Cloudflare API key.

Who is IdentiClaw for?

Developers, teams and enterprises running autonomous agents, and anyone who wants to try an agent without exposing their passwords.

The KYA newsletter

Be first to know what agents do next.

AI agents change something every week. Once a month, we’ll send you what’s new in agent identity, KYA‑OS and the fixes that actually work, so you’re always a step ahead. No spam, ever.

By subscribing, you agree to our Privacy Statement. Unsubscribe in one click.