> ## Documentation Index
> Fetch the complete documentation index at: https://docs.blockgence.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Guard

> Active monitoring of your own wallets or an attacker's, with alerts in the application, by signed webhook and by email.

> [!NOTE]
>
> The application interface is in Spanish. Screens, tabs and buttons are referred to by their name in the application,
> followed by their English meaning.

Guard watches an address and reports its activity, block by block. It has two uses:

> [card] Your own wallets
>
> Operational and treasury wallets: a high-value outflow, an outflow the wallet did not sign, a new approval, a
> delegation or a risky counterparty.

> [card] An attacker's wallet
>
> Every outflow and its destination: an exchange (the right moment to request a freeze), a bridge, a mixer, or a new
> account that Guard can follow automatically.

It is available at **app.blockgence.com → Guard**. The tabs are **Alertas** (alerts) and **Wallets**, plus **Avisos**
(notifications: webhook and email settings) and **Protocolos** (protocols: Watchtower).

![Watched wallets in Guard](https://mintcdn.com/blockgence/jIDqBW86tOU_Ini-/images/products/guard/wallets.png?fit=max\&auto=format\&n=jIDqBW86tOU_Ini-\&q=85\&s=5299e405c6f439a568308cf126492c43)

Watched wallets with synthetic data.

## Watch an address

**1. Enter the address and the network**

In **Wallets**, type the address and select the network. The top of the screen shows which networks Guard watches and
whether they are up to date.

**2. Choose the purpose**

**Una wallet propia** (your own wallet) or **La wallet de un atacante** (an attacker's wallet). Each purpose has its
default rules, which can be adjusted later.

**3. Share the watch (optional)**

A shared watch is visible to the whole organization, and anyone who can see it can manage its alerts.

Guard watches **from the moment the address is added, not retroactively**. The detail of each watch shows the block from
which it is watched, along with pauses and service interruptions. For earlier activity, use the trace in
[Investigation](https://docs.blockgence.com/en/products/investigation.md).

A watch can also be started from other parts of the application: the **vigilar** (watch) link on an address in
[Monitor](https://docs.blockgence.com/en/products/monitor.md), the last step of incident mode (which adds the attacker's accounts) and the copilot,
upon request.

## Recorded activity

Guard records the transaction the address signs (even if it reverts), the native coin with its internal movements,
ERC-20 tokens, ERC-721 and ERC-1155 NFTs, approvals (including Permit2), EIP-7702 delegations (even when submitted by a
third party) and the contracts it creates. Each event is valued in USD and includes the counterparty's label.

**Noise** is recorded, marked, and does not trigger alerts: spam tokens, zero-value transfers (typical of address
poisoning), spoofed senders and negligible amounts.

## Rules

| Rule | Condition | Severity | Own | Attacker |
| - | - | - | - | - |
| High-value outflow | A priced outflow above a threshold (default: US\$10,000) | High | Yes | No |
| High-value inflow | A priced inflow above a threshold (default: US\$1,000) | Medium | No | Yes |
| Any outflow | Every signed transaction and every outflow of funds | Medium | No | Yes |
| Unsigned outflow | A third party uses an approval granted by the address | Critical on an own wallet | Yes | Yes |
| Approval | A new approval (not revocations) | High; critical if the spender is risky | Yes | Yes |
| Delegation | An EIP-7702 delegation | Critical | Yes | Yes |
| Risky counterparty | Funds to or from a sanctioned, darknet or scam address, a mixer or a known attacker | Critical outgoing, high incoming | Yes | Yes |
| Exchange | An outflow to an exchange | Critical | No | Yes |
| Bridge | An outflow to a bridge | High | No | Yes |
| Contract created | The address deploys a contract | High | No | Yes |
| Follow the hops | Also watch the new accounts that receive its funds | High | No | Disabled by default |

A transaction generates **a single alert** per watch, with the most severe rule and the list of every rule it triggered.

### Follow the hops

On an attacker's wallet, **Seguir los saltos** (follow the hops) makes Guard also watch the accounts that receive its
funds, starting at US\$1,000 accumulated per destination. This prevents the attacker from evading the watch by changing
addresses. The derived watch inherits the rules, follows its own hops up to a depth of 2 (10 derived watches per tree,
adjustable) and counts toward the plan limit.

> [!NOTE]
>
> Only regular accounts without a label are followed. Exchanges, bridges, routers, mixers, contracts and EIP-7702
> delegated accounts are not followed, and each stop is shown with its reason. An account that received the attacker's
> funds does not necessarily belong to the attacker: the derived watch can be deleted.

## Manage alerts

Each alert states what happened and where the funds went, with the transaction and the block. An alert can be
**reconocida** (acknowledged), **asignada** (assigned), **resuelta** (resolved) with a reason or **descartada**
(dismissed), and every change is kept in its history. **Abrir un caso en Investigación** (open a case in Investigation)
turns the alert into a case with the transaction's graph, without using credits. If a network reorganization removes the
transaction, the alert is retracted (not deleted) and is reactivated if the transaction is included again.

![Guard alert detail](https://mintcdn.com/blockgence/jIDqBW86tOU_Ini-/images/products/guard/alert.png?fit=max\&auto=format\&n=jIDqBW86tOU_Ini-\&q=85\&s=df1aa0f8a9e7b0c2409b6536aafc4c53)

Alert detail with synthetic data.

## Notifications outside the application

> [tab] Signed webhook
>
> A POST request with one JSON payload per alert and event, signed with HMAC-SHA256 according to **Standard Webhooks**,
> with one secret per destination that can be rotated. The events are: alert created, retracted, reactivated and updated
> by a person (acknowledged, assigned, resolved, dismissed, reopened or commented); the last one states who updated the
> alert and what changed. The scope can be one watch, your own
> watches or the organization's shared watches. Failed deliveries are retried for approximately 15 hours. From
> **Avisos** (notifications) you can send a test and review the deliveries.

> [tab] Email digest
>
> One digest per person, never one email per alert. By default, it includes only high and critical alerts. It does not
> include updates made by people.

> [!WARNING]
>
> The webhook reports alerts and their changes, not every movement of the watched address.

## Watchtower: a protocol's contracts

A protocol registers its contracts in **Protocolos** (protocols) and Guard watches them with specific rules: a
high-value net withdrawal per transaction in USD (default: US\$1,000,000), a change of owner, admin or administrator role,
a proxy upgrade, a pause, and two-step owner or admin transfers (it alerts on the proposal before it is accepted). It also
reads the events the protocol declares and, if its timelock or governor is registered, reports what a scheduled call
would change before it is executed.

> [!NOTE]
>
> Watchtower is disabled by default. If it is not enabled, the **Protocolos** tab states so.

## API usage

The same features are available under `/v1/guard/*`, with an API key **assigned to a person**; a key without an owner
receives a 403 response. Guard does not use credits: the limit is the number of watches in the plan.

```bash Watch an address theme={null}
curl -X POST https://app.blockgence.com/v1/guard/watches \
  -H "X-API-Key: $BKG_KEY" -H "Content-Type: application/json" \
  -d '{"chain": "ethereum", "address": "0x2222222222222222222222222222222222222222", "name": "Treasury", "purpose": "own"}'
```

```bash Active alerts theme={null}
curl https://app.blockgence.com/v1/guard/alerts -H "X-API-Key: $BKG_KEY"
```

Each route is detailed in the [API reference](https://docs.blockgence.com/en/api/reference.md).

## Scope and limitations

* **Monitoring starts when the address is added**, not retroactively.
* Ethereum is watched block by block. On Base, Optimism, Arbitrum, BSC and Polygon, when enabled, coverage is narrower:
  monitoring relies on the events of the watched addresses, and anything that cannot be located is declared on each
  event.
* The **false alert rate** on real wallets is not measured.
* Watchtower only observes what the contract emits. The owner of a ProxyAdmin, the owner of a Safe or a beacon upgrade
  are not covered unless those contracts are registered as well, and the response declares it.
* Guard does not block transactions or revoke permissions: it reports.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.