-
OPEN SOURCE · DEVELOPER-FIRST
Sign your agents.
Prove they
haven't changed.
nxtlinq Attest is a lightweight, cryptographic signing and verification system for AI agents. Prevent tampering, enforce declared scope, enable revocation — and optionally anchor to the nxtlinq transparency registry. No friction. Works locally by default.
NXTLINQ-ATTEST · BASH
# Install
$ npm install -g @nxtlinq/attest
# Initialize your agent project
$ nxtlinq-attest init
Initialized nxtlinq attest in /path/to/your/agent
- nxtlinq/private.key (do not commit)
- nxtlinq/public.key
- nxtlinq/agent.manifest.json
Next: edit nxtlinq/agent.manifest.json — set name, version, and scope. Then run: nxtlinq-attest sign
# Sign your agent release
$ nxtlinq-attest sign
Signed manifest and artifact.
contentHash: fcb5c0888ade248b...
artifactHash: a9cf23b51c4bfea1...
artifactFileCount: 428
signature: nxtlinq/agent.manifest.sig
# Verify before running
$ nxtlinq-attest verify
Verification passed.
name: my-agent version: 1.0.0
scope: agent:my-agent
artifactFileCount: 428
-
OPENCLAW COMMUNITY
Built for where agents are actually being built.
OpenClaw is where developers prototype, share, and deploy AI agents at scale. That's exactly where supply-chain risk starts. nxtlinq Attest ships as a native OpenClaw skill — so every agent built in the community can be signed, verified, and trusted before it reaches production. If you build on OpenClaw, Attest is already in your workflow.
10K+
OPENCLAW
DEVELOPERS
1st
AGENT SIGNING
SKILL ON OPENCLAW
60s
TIME TO SIGN
YOUR FIRST AGENT
HOW IT WORKS ———
Three steps. Zero friction.
STEP 02 —————————
Sign your agent release
Cryptographically sign the agent artifact, producing nxtlinq/agent.manifest.sig that travels with the agent.
nxtlinq-attest sign
STEP 03 —————————
Verify before execution
Before running any agent, validate its signature integrity and confirm scope consistency. This can be enforced in CI to ensure no modified or unsigned agent is deployed without a clean attestation.
nxtlinq-attest verify
OPENCLAW INTEGRATION ———
The community where
agents are born.
OpenClaw is the agentic AI operating platform where developers prototype, collaborate, and ship agents at scale. Its Room-based architecture — where humans, agents, and tools share a governed context — is precisely where unverified agents create supply-chain risk. nxtlinq Attest is the first signing standard built natively into this ecosystem.
As an OpenClaw skill, Attest is a one-command install for any developer already building there. You don't leave your workflow. You don't set up new infrastructure. You sign your agents exactly where you build them — and the attestation travels with every deployment, fork, and share.
01
Install as an OpenClaw skill
One command adds Attest to your OpenClaw Room. No new accounts, no new platforms. Your agents, your workflow — now signed.
02
Sign before you share
Every agent you share from OpenClaw carries a verifiable signature. Other developers can confirm your agent hasn't been tampered with before running it in their Rooms.
03
Scope drift detection at the community level
When an agent's declared scope doesn't match its verified manifest, Attest flags it. OpenClaw becomes the first agent marketplace with built-in scope enforcement.
04
Path to enterprise trust
Agents prototyped on OpenClaw and signed with Attest carry a verifiable provenance chain into production. The trust you establish in the community follows the agent upstream.
DEVELOPER WORKFLOW ———
From prototype to
signed production agent.
Build your agent in a Room
OPENCLAW
Prototype in OpenClaw's collaborative agentic environment. Define your agent's tools, permissions, and behavior.
Sign with one command
ATTEST SKILL
Run the Attest skill inside your Room. Manifest generated, scope bound, cryptographic signature produced.
nxtlinq-attest sign
Graduate to enterprise identity
NXTLINQ NXTID
Agents with Attest provenance can be onboarded to the full nxtlinq HIT/AIT governance fabric — carrying their signed identity chain from OpenClaw prototype to enterprise production deployment.
TRUST MODEL ———
Progressive trust. Zero forced friction.
Level 0 — Self-Signed
Local integrity
Signature verified locally. Integrity guaranteed for your own use. No external dependencies. Perfect for active development and OpenClaw prototyping.
Level 1 — Logged
Publicly recorded
Manifest hash published to the nxtlinq registry. Publicly recorded and not revoked. Shareable with other OpenClaw developers and downstream consumers with full verifiability.
Level 3 — Enterprise Bound
Full provenance chain
Organization-linked provenance and auditable HIT/AIT chain via nxtlinq nxtID. Every agent action is attributable to a verified human origin and governed identity fabric.
FUTURE PHASE
Level 2 — Issuer Verified
Recognized issuer
Signed by a recognized issuer in the nxtlinq registry. Signals organizational accountability — ideal for open-source maintainers, platform vendors, and agent publishers with downstream users.
DISTRIBUTION ———
Three surfaces. One standard.
CLI
The core tool. Install globally via npm. Sign, verify, publish, and revoke from any terminal. Works locally with zero external dependencies at Level 0.
npm install -g @nxtlinq/attest
AVAILABLE NOW
GitHub Action
Integrate agent verification into your CI/CD workflow.
Use nxtlinq-attest verify in a GitHub Actions workflow to ensure only signed agents are merged.
Example: run verification as a step in your pull request pipeline
AVAILABLE NOW
OpenClaw Skill
Native skill wrapper for OpenClaw's Room-based agent environment. Install Attest into any Room in one command. Sign agents exactly where you build them — no context switching required.
skill install nxtlinq-attest