Sentinel Alpha is live on Base

Check one agent action before it spends, calls a tool, or touches an asset

BeezShield returns allow/review/block decision support, checked vs not-checked boundaries, and a path to an audit-ready Trust Receipt.

Run Free Action Preview Request $50 Trust Receipt Pilot

Agentic commerce checkout

Verify merchant identity, check product price labels, and validate order details before allowing purchase approval.

API/tool call governance

Govern schema parameters, payload destinations, and execution limits to stop malicious tool execution.

Data/API credit purchases

Review vendor billing endpoints and credit spend before agents purchase API capacity or datasets.

Workflow execution

Gate multi-step automations with explicit checked vs not-checked boundaries before production runs.

Onchain contract/asset action

Pre-verify contract targets and asset movements on Base or mainnet before agents sign or broadcast.

x402scan: Registered

Agentic.Market: Validator passed (not listed)

Pay.sh: Not submitted

Ampersend: Not submitted

Sentinel + Trust Receipt Pipeline

Verification at every stage of the autonomous agent lifecycle.

PRE-EXECUTION

Sentinel Risk Decision

Before an agent performs any wallet automation or contract execution, Sentinel assesses the transaction target's threat profile, blocking execution on high-risk signatures or unverified contract entities.

Query pre-execution risk score via POST
0.02 USDC per-request basic lane pricing
Real-time threat classification

POST-EXECUTION

Trust Receipt Packet

Once execution completes, the agent compiles a cryptographically signed Trust Receipt. The receipt acts as an auditable verification packet of the agent's pre-execution inputs and post-execution logs.

Cryptographically signed receipts
Redacted logs & payload details
Auditable by third-party protocols

What this proves

The Sentinel + Trust Receipt pipeline documents that a specific agent performed pre-execution contract verification before onchain action. It links point-in-time threat signals to a redacted proposed-action summary and optional post-execution hash refs — not proof that an onchain transaction matched a validated target.

What this does not prove

This pipeline is not a security guarantee. It does not prove the logical correctness of the target contract's code, nor does it guarantee protection against zero-day exploits. It is not an official endorsement or partnership with the target protocols or base infrastructure providers.

Learn More & Request Trust Receipt Pilot

Use Cases

Agent execution boundaries that require trust infrastructure.

Before agent contract execution

Problem: Agents cannot blindly sign payloads to arbitrary contracts.

Action: Request a risk-score for the target before execution.

Note: Reduces risk but does not guarantee total security.

Before wallet automation

Problem: Automated wallets are prime targets for drainers.

Action: Gate transactions behind a Sentinel approval threshold.

Note: Adds a layer of defense; not a replacement for good ops.

Before onchain task delegation

Problem: Handing off tasks to unverified smart contracts is dangerous.

Action: Validate the target's threat class and history.

Note: Evaluation based on point-in-time metrics.

Before marketplace tool invocation

Problem: 3rd party agent tools can act as honeypots.

Action: Confirm tool contract reputation before calling.

Note: Does not audit logic, only infrastructure signals.

Integration Paths

Designed for autonomous execution. Documented for human developers.

FOR AGENTS

Automated Execution Pipeline

Integrate pre-execution decision checks directly into autonomous agent loops.

POST Endpoint: /contracts/risk-score
x402 Payment Loop: Reads 402 status, micro-payment on Base
JSON Payloads: Compact input address, output action keys

View API Spec

FOR HUMANS

Developer Tooling & Auditing

Everything human builders need to audit, configure, and initialize BeezShield.

Detailed Docs: API Reference & guides
npm SDK: Zero-dependency package @beezshield/sentinel
GitHub Repository: Open-source SDK & templates chox-cell/Sentinel-Alpha
Registry Page: Public discoverability x402scan record

Read Documentation

How agents pay

Machine-to-machine commerce using the x402 protocol.

BeezShield uses x402 HTTP status codes to request payment. Agents read the 402 Payment Required response, submit a transaction, and retry with the payment header.

Basic lane price 0.02 USDC
Network Base

We are registered as a payable x402 resource on x402scan. This is a directory listing only, and does not imply partnership or official x402 integration.

1. Unpaid 402 discovery curl

curl -s -X POST https://api.beezshield.com/contracts/risk-score | jq .

2. POST sample body

{
  "contract_address": "0x1111111111111111111111111111111111111111",
  "chain": "base"
}

Developer Preview

Build with Sentinel Alpha

Pre-execution decision engine for agents. Developer quickstart: The Sentinel Alpha SDK (@beezshield/sentinel) is available on npm (version 0.1.0). Read documentation on how to use functions like createSentinelClient(), scoreContract(), and decideBeforeExecution(). Install it via: npm install @beezshield/sentinel. Upcoming integration: AgentKit integration: upcoming integration. View the AgentKit-style example in our repo.

AgentKit integration: upcoming integration

# Discovery (returns 402 Payment Required)
curl -X POST https://api.beezshield.com/contracts/risk-score

# Paid request
curl -X POST "https://api.beezshield.com/contracts/risk-score" \
  -H "Content-Type: application/json" \
  -H "X-SENTINEL-LANE: basic" \
  -H "X402-PAYMENT: tx:YOUR_TX_HASH" \
  -d '{"contract_address":"0x...","chain":"base"}'

import { createSentinelClient } from "@beezshield/sentinel";

const client = createSentinelClient({
  x402Payment: "tx:YOUR_TX_HASH"
});

const decision = await client.decideBeforeExecution({
  contract_address: "0x...",
  chain: "base",
});

# Python integration
# SDK method decodes the pre-execution risk score
result = client.decideBeforeExecution({
  "contract_address": "0x...",
  "chain": "base"
})

Add to your bot npm: @beezshield/sentinel OpenAPI JSON Live API Endpoint

Pricing

x402 lanes on Base USDC — same tiers on pricing.html and x402 guide.

BASIC

0.02 USDC

per query

EXECUTIVE

0.05 USDC

per query

PREMIUM

0.10 USDC

per query

PRIORITY

0.15 USDC

per query

Unpaid POST /contracts/risk-score returns an x402 payment challenge. Settlement is explicit in your client — not automatic in the SDK.

Trust & Proof

Verifiable infrastructure presence.

What this does not prove

BeezShield is a decision support tool. Our presence in directories or verifiable identity does not equate to a security guarantee, nor does it imply an official partnership or endorsement by the x402 protocol.