Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 11 additions & 7 deletions .cursor/rules/product-agent-wallet.mdc
Original file line number Diff line number Diff line change
Expand Up @@ -16,8 +16,8 @@ alwaysApply: false

- **server-wallet** — private keys are held securely in a TEE-backed environment.
- **Bring your own wallet** — user supplies a BIP-39 mnemonic. CLI flag: `byok`.
- **Guard Mode (Recommended)** — trading mode (`--mode guard`). Designed for everyday traders. Enforces security check, network allowlist, token recipient allowlist, address allowlist, and rolling 24h outflow limit. 2-factor authentication on policy violations, malicious transactions, and raising outflow limit.
- **Beast Mode** — trading mode (`--mode beast`). Designed for power users. Security check guardrail only. 2-factor authentication on malicious transactions and risky contracts.
- **Guard Mode (Recommended)** — trading mode (`--mode guard`). Designed for everyday traders. Enforces threat scanning, network allowlist, token recipient allowlist, address allowlist, and rolling 24h outflow limit. 2-factor authentication on policy violations, malicious transactions, and raising outflow limit.
- **Beast Mode** — trading mode (`--mode beast`). Designed for power users. Threat scanning guardrail only. 2-factor authentication on malicious transactions and risky contracts.

Users choose wallet and trading modes during interactive `mm init` or by prompting their agent.
Mark Guard Mode as **Recommended**; do not describe modes as opt-in, default, or preferred in other ways.
Expand All @@ -29,13 +29,17 @@ CLI flags use `server-wallet`, `byok`, `guard`, and `beast`.

Use these terms consistently when describing the security stack:

- **Transaction Protection** — guarantee for transactions deemed safe (up to $10,000 loss coverage).
Frame as backing for "security by default," not as a standalone scanning feature. Link eligibility
- **Threat scanning** — powered by Blockaid; production-tested across millions of MetaMask
transactions. Malicious transactions get auto-bounced. Do not conflate with Transaction Shield.
- **Transaction Protection** — guarantee for eligible transactions deemed safe (up to $10,000/month
loss coverage). Frame as optional coverage backing, not as a scanning feature. Link eligibility
and terms to
[Transaction Shield support](https://support.metamask.io/manage-crypto/transactions/transaction-shield/).
- **Transaction Shield** — automated security checks for malicious contracts and scams. Link to
[Transaction Shield support](https://support.metamask.io/manage-crypto/transactions/transaction-shield/).
Do not lead with vendor names.
- **Transaction Shield** — MetaMask subscription pairing Transaction Protection with priority support.
Link to
[Transaction Shield support](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)
for subscription details, eligibility, and terms. Do not describe Transaction Shield as a security
or threat-scanning feature.
- **Smart Transactions** — optimizes how trades land onchain (fewer fails, better gas, built-in MEV
protection). Link to
[Smart Transactions support](https://support.metamask.io/manage-crypto/transactions/smart-transactions/).
Expand Down
21 changes: 12 additions & 9 deletions agent-wallet/README.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -33,23 +33,26 @@
## MetaMask safety-checks every transaction before it lands

Security by default, not configuration.
Supported EVM transactions pass through a mandatory 3-step security pipeline.
Backed by Transaction Protection: transactions through Agent Wallet deemed safe are guaranteed
against loss up to $10,000.\*
Supported EVM transactions pass through a mandatory 3-step pipeline before they land onchain.

1. **Transaction simulation**: What will this transaction actually do? Balance changes, approvals,
and gas are surfaced before signing.
2. **[Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)**:
Automated security checks flag malicious contracts and scams; flagged transactions require your
approval before they execute.
2. **Threat scanning**: Powered by Blockaid and production-tested across millions of MetaMask

Check warning on line 40 in agent-wallet/README.mdx

View workflow job for this annotation

GitHub Actions / Spelling (.md)

[vale] reported by reviewdog 🐶 [Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary. Raw Output: {"message": "[Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary.", "location": {"path": "agent-wallet/README.mdx", "range": {"start": {"line": 40, "column": 36}}}, "severity": "WARNING"}

Check warning on line 40 in agent-wallet/README.mdx

View workflow job for this annotation

GitHub Actions / Spelling (.mdx)

[vale] reported by reviewdog 🐶 [Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary. Raw Output: {"message": "[Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary.", "location": {"path": "agent-wallet/README.mdx", "range": {"start": {"line": 40, "column": 36}}}, "severity": "WARNING"}
transactions. Malicious transactions get auto-bounced. Flagged transactions require your approval
before they execute.
3. **[Smart Transactions](https://support.metamask.io/manage-crypto/transactions/smart-transactions/)**:
Smart transaction execution with built-in MEV protection, fewer fails, and better gas where
supported on the target chain.

Eligible transactions deemed safe are backed by
[Transaction Protection](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)
coverage up to $10,000/month.\*

:::note

Learn more in [Architecture](reference/architecture.md). See [Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)
for eligibility, coverage limits, and terms.
Learn more in [Architecture](reference/architecture.md). See
[Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/) for
subscription details, eligibility, coverage limits, and terms.

:::

Expand Down Expand Up @@ -96,7 +99,7 @@
href: '/agent-wallet/reference/architecture',
title: 'Architecture',
description:
'Wallet modes, Transaction Protection, Transaction Shield, Smart Transactions, and 2FA.',
'Wallet modes, threat scanning, Transaction Protection, Smart Transactions, and 2FA.',
},
{
href: '/agent-wallet/guides/trade-perpetuals',
Expand Down
4 changes: 2 additions & 2 deletions agent-wallet/quickstart.md
Original file line number Diff line number Diff line change
Expand Up @@ -89,10 +89,10 @@ During `mm init`, choose a wallet mode and, for server-wallet, a trading mode.

#### Trading mode (server wallet only):

- **Guard Mode (Recommended)**: designed for everyday traders. Enforces security checks, network
- **Guard Mode (Recommended)**: designed for everyday traders. Enforces threat scanning, network
and recipient allowlists, address allowlists, and a rolling 24-hour outflow limit. Transactions
outside your policy limits require 2-factor authentication before they execute.
- **Beast Mode**: designed for power users. Keeps the security check guardrail only. Malicious
- **Beast Mode**: designed for power users. Keeps the threat scanning guardrail only. Malicious
transactions and risky contracts are blocked and surfaced for 2-factor authentication approval.

See [Trading modes](reference/architecture.md#trading-modes) for guardrails and approval conditions.
Expand Down
26 changes: 18 additions & 8 deletions agent-wallet/reference/architecture.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@
server-wallet,
polling,
Transaction Protection,
Transaction Shield,
threat scanning,
Smart Transactions,
2FA,
]
Expand Down Expand Up @@ -61,13 +61,23 @@
Before a transaction executes, the CLI simulates it to surface reverts, unexpected state changes,
and other failures early.

## Transaction Shield
## Threat scanning

[Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)
runs automated security checks on each transaction, including malicious contracts and scams.
Threat scanning is powered by Blockaid and production-tested across millions of MetaMask

Check warning on line 66 in agent-wallet/reference/architecture.md

View workflow job for this annotation

GitHub Actions / Spelling (.md)

[vale] reported by reviewdog 🐶 [Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary. Raw Output: {"message": "[Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary.", "location": {"path": "agent-wallet/reference/architecture.md", "range": {"start": {"line": 66, "column": 31}}}, "severity": "WARNING"}

Check warning on line 66 in agent-wallet/reference/architecture.md

View workflow job for this annotation

GitHub Actions / Spelling (.mdx)

[vale] reported by reviewdog 🐶 [Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary. Raw Output: {"message": "[Consensys.Spelling] Did you really mean 'Blockaid'? Ignore this alert if this is a false positive, or ask Cursor to add the term to the Vale dictionary.", "location": {"path": "agent-wallet/reference/architecture.md", "range": {"start": {"line": 66, "column": 31}}}, "severity": "WARNING"}
transactions.
Malicious transactions get auto-bounced.
When a transaction is flagged, it requires your approval before it executes.
You receive details in the CLI output and through the approval flow.

## Transaction Protection

Eligible transactions deemed safe are backed by Transaction Protection coverage up to
$10,000/month.
[Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/) is
MetaMask's subscription that pairs Transaction Protection with priority support.
See [Transaction Shield](https://support.metamask.io/manage-crypto/transactions/transaction-shield/)
for subscription details, eligibility, coverage limits, and terms.

## Smart Transactions

[Smart Transactions](https://support.metamask.io/manage-crypto/transactions/smart-transactions/)
Expand All @@ -93,7 +103,7 @@

**Guardrails**

- Security check
- Threat scanning
- Network allowlist
- Token recipient allowlist
- Address allowlist
Expand All @@ -114,7 +124,7 @@

**Guardrails**

- Security check
- Threat scanning

**Approval required for**

Expand All @@ -140,8 +150,8 @@
When you submit a signing or transaction request in server-wallet mode:

1. The CLI submits the request to the wallet service.
2. The service may simulate the transaction, run Transaction Shield, and evaluate policies.
3. If policy requires 2-factor authentication or Transaction Shield flags the transaction, the job
2. The service may simulate the transaction, run threat scanning, and evaluate policies.
3. If policy requires 2-factor authentication or threat scanning flags the transaction, the job
enters an `AWAITING_MFA` state until you approve via MetaMask Mobile or email.
4. The CLI returns a `pollingId` unless you pass `--wait`.

Expand Down
4 changes: 2 additions & 2 deletions agent-wallet/use-the-cli-directly.md
Original file line number Diff line number Diff line change
Expand Up @@ -78,11 +78,11 @@ Optionally encrypt the mnemonic at rest with `mm wallet password set`.
| Guard Mode (Recommended) | `--mode guard` | Designed for everyday traders. Transactions outside your policy limits require 2FA approval. |
| Beast Mode | `--mode beast` | Designed for power users. Malicious transactions are still blocked and surfaced for 2FA approval. |

Guard Mode enforces security checks, network and recipient allowlists, address allowlists, and a
Guard Mode enforces threat scanning, network and recipient allowlists, address allowlists, and a
rolling 24-hour outflow limit.
2FA is required for malicious transactions, allowlist violations, and raising your outflow limit.

Beast Mode keeps only the security check guardrail.
Beast Mode keeps only the threat scanning guardrail.
2FA is required for malicious transactions and risky contracts.

See [Trading modes](reference/architecture.md#trading-modes) for the full guardrail and approval
Expand Down
Loading