Reusable AgentSkill · Pharos Atlantic

The Skill your agent calls
before it moves value.

Name an address. Sentinel judges it.

Sentinel is the pre-action risk Skill an onchain agent calls before it signs. It reads the target cold (bytecode, storage, ownership, token state) and returns a safe / caution / dangerous verdict with reasons.

The gate below is that Skill running live. Drop in an address (try the Backdoor) and you see exactly what an agent sees when it calls Sentinel.

RISK GATE · read-only cast call / code / storage
Clean token Zero-supply trap Upgradeable proxy Backdoor

        
SAFETY SCORE

Findings

688689
Atlantic chainId
9
on-chain signals
0
keys held
read-only
never signs
How it works

Follow one check through the gate.

Every verdict is built from read-only Foundry cast reads of the live chain — no indexer, no third-party API.

01

Read the chain

Sentinel pulls the target's bytecode, proxy storage slots, and token calls — all read-only, all through Foundry.

02

Detect the signals

It walks the opcodes and proxy slots for the patterns that actually drain wallets — not surface heuristics. Each one found subtracts from the safety score.

03

Score it

Every check starts at a perfect 100; each signal subtracts from it, so multiple small risks pull the safety score down.

04

The verdict

A band turns the score into a clear decision — safe, caution, or dangerous — an agent can act on directly.

05

Always the reasons

Never just a number. Every verdict carries the human-readable why — the audit trail behind the call.

06

Gate the action

The verdict becomes an action — approve or block, plus a de-risked size — so value moves only when it's safe to.

The gate is open

Check an address. Get the verdict.

Read-only, keyless, reproducible — the most-called primitive in any on-chain agent stack.

Run a check ▸