Introducing the AXIS OpenClaw Skill: Instant Trust Verification for Any AI Agent

The official AXIS OpenClaw Skill gives any OpenClaw-compatible agent framework immediate access to trust verification, score lookups, agent registration, and behavioral event reporting — with no custom integration required. Drop in the SKILL.md, and your agent can check a counterparty's trust tier before it transacts, delegates, or shares data.

By Leonidas Esquire Williamson — March 13, 2026

The AXIS platform now ships an official OpenClaw skill — a single SKILL.md file that gives any OpenClaw-compatible AI agent framework immediate access to the full AXIS trust infrastructure. No custom HTTP client, no API key wiring, no schema mapping. Drop in the file and your agent gains the ability to verify trust, look up scores, register itself, and report behavioral events.

What Is an OpenClaw Skill?

OpenClaw is a skill-based capability framework for AI agents. A skill is a structured Markdown file that describes a set of capabilities, their inputs, their outputs, and the exact API calls required to exercise them. An OpenClaw-compatible agent runtime reads the skill file at startup and exposes its capabilities to the agent's reasoning loop — the agent can then decide when and how to use them based on context.

The AXIS OpenClaw Skill defines five capabilities:

Trust Verification is the most commonly used capability. Given any AUID, the agent can retrieve the counterparty's T-Score (0–1000), C-Score (0–1000), trust tier (T1–T5), and registration date. This call is public — no authentication required — which means an agent can verify a counterparty's trust standing before it has established any relationship with AXIS itself.

Agent Registration allows an agent to register itself with the AXIS network and receive a permanent AUID. Registration requires authentication via API key and takes a display name, agent class, and home registry as inputs. Once registered, the agent's AUID can be shared with counterparties and used as a stable identity across all interactions.

Score Retrieval provides authenticated access to detailed T-Score and C-Score breakdowns, including tier history, event counts, and score trajectory. This is useful for agents that need to monitor their own standing or audit their score changes over time.

Event Reporting is how agents contribute to the trust network. After completing an interaction — a successful task delegation, a payment, a data exchange — the agent reports the outcome as a behavioral event. The event type, score impact, and a brief description are submitted to AXIS, which applies credibility weighting and routes the event through the appropriate verification pathway (direct application or Pending Verification Queue, depending on the event type).

API Key Management allows agents to list and create API keys programmatically, which is useful for agents that manage their own credentials or rotate keys on a schedule.

The Trust Decision Framework

The skill includes a built-in Trust Decision Framework — a set of tier-based rules that tell the agent what actions are appropriate given a counterparty's trust standing:

These rules are embedded directly in the skill file, so the agent's reasoning loop can apply them without any additional configuration.

Installation

Installing the AXIS OpenClaw Skill takes three steps.

First, download the skill file from the AXIS platform:

```bash

curl -o AXIS-OpenClaw-SKILL.md \

https://www.axistrust.io/docs#openclaw

```

Second, place the file in your agent's skills directory (the exact path depends on your OpenClaw runtime):

```bash

cp AXIS-OpenClaw-SKILL.md ~/.openclaw/skills/axis-trust.md

```

Third, if you want to use authenticated capabilities (score retrieval, event reporting, API key management), set your AXIS API key as an environment variable:

```bash

export AXIS_API_KEY="axk_your_api_key_here"

```

That is the entire installation. The next time your agent starts, it will have access to all five AXIS capabilities.

A Trust Check in Practice

Here is what a trust verification call looks like from the agent's perspective. The agent wants to verify a counterparty before delegating a task:

```bash

Check counterparty trust standing (no auth required)

curl -s "https://www.axistrust.io/api/trpc/agents.getByAuid" \

-G --data-urlencode 'input={"json":{"auid":"axis:autonomous.registry:T3:7k2m3n4p:a3f7"}}' | \

python3 -c "

import sys, json

data = json.load(sys.stdin)

agent = data['result']['data']['json']

print(f'Name: {agent["name"]}')

print(f'T-Score: {agent["tScore"]}')

print(f'Tier: {agent["tier"]}')

print(f'C-Score: {agent["cScore"]}')

"

```

The response arrives in under 100 ms. The agent reads the tier, applies the Trust Decision Framework, and proceeds — or declines — accordingly. No human in the loop, no manual lookup, no guesswork.

Reporting an Interaction Outcome

After a successful task delegation, the agent reports the outcome:

```bash

Report a successful task delegation (requires API key)

curl -s -X POST "https://www.axistrust.io/api/trpc/trust.addEvent" \

-H "Authorization: Bearer $AXIS_API_KEY" \

-H "Content-Type: application/json" \

-d '{"json":{"agentId":42,"eventType":"TASK_COMPLETED","category":"delegation","scoreImpact":15,"description":"Delegated data analysis task completed on time and within spec"}}' | \

python3 -c "import sys,json; r=json.load(sys.stdin); print('Event recorded:', r['result']['data']['json']['id'])"

```

The event is processed, credibility-weighted by the reporter's trust tier, and applied to the counterparty's T-Score. If the event type requires dual-party verification, it enters the Pending Verification Queue and the counterparty is notified via webhook.

Backend-Verified

The AXIS OpenClaw Skill was audited against the live backend before release. Five bugs in the original draft were identified and corrected: two wrong field names (class → agentClass, name → label), two missing agent class values (research and service), an undocumented integer requirement for agentId, and a stale development URL replaced with the canonical production domain. All curl examples were updated to use the correct tRPC v11 superjson input wrapper — the previous format returned BAD_REQUEST errors.

The corrected skill file is the version available for download. The SHA-256 hash of the current release is available on the [Changelog](/changelog) page.

Get Started

The skill file is available for download from the [Docs page](/docs#openclaw) and from the [OpenClaw Skill](/docs#openclaw) link in the navigation bar. If you have questions about integration, the [API Explorer](/api-explorer) provides an interactive reference for every AXIS endpoint.