Skip to main content

​
What it is

Protocol Guard is the monitoring surface. You tell it what a wallet is about to do — borrow, vote, hit an API, touch the treasury — and it returns whether that action clears a contextual floor, plus any alerts worth firing. It does not replace your risk engine. It is one input, with a recommended default threshold per context that you can override. Live UI: trustgated.xyz/protocol-guard Pricing: free. Registration exists only to save alert channels and rules. Checks work anonymously without it.

​
Check

context is one of: borrow, vote, governance_vote, dex_swap, api_execution, treasury_control, generic. The response includes allowed, walletScore, walletTier, confidence, flags, a contextualThreshold, a trustSurfaceArea (monitor / review / immediate), and any alerts. These are defaults, not TrustGate policy. Override them per request or per registration.

​
Trust Surface Area

When you pass economicReachUsd, capitalAccess, and (when relevant) a vote cluster, Guard computes a surface-area number used only to prioritize alerts:
  • monitor — low potential damage
  • review — look before the window closes
  • immediate — low trust × high reach
Higher surface means more potential damage if the actor is adversarial, not “this wallet is guilty.”

​
Registration and alerts

Register a protocolName, optional contactEmail, channels (Discord webhook, Telegram, email, onchain event flag), and rules:
  • lending_low_trust_borrow
  • dao_coordinated_vote
  • wallet_score_floor
  • token_flag_watch
A manageToken is shown once at create. You need it to GET-by-id or PATCH.
Subscriptions and alert state are stored in process memory on the current deployment. They do not survive a restart and are not shared across isolates. Fine for pilots. Not a durable billing or paging system.

​
Build it

Protocol Guard integration

Check body, contexts, registration, and the free-anonymous path.

Gating instead

If you need a signed attestation, not an alert.