LighthouseAgent

Transaction firewall for AI agents on @pharos_network. Every tx simulated, policy-checked & logged on-chain before signing. Built on Anvita Flow.

[LighthouseAgent @Lighthouse_xyz]

Coming soon to Lighthouse Agent 🗼
Swaps + liquidity on @pharos_network — checked by the agent itself before signing.
Every swap and LP move passes the built-in firewall: router allowlist, exact approvals only, slippage & price-impact bounds, on-chain verdict log.
The agent that trades — and audits its own trades.

[LighthouseAgent @Lighthouse_xyz]

Where Lighthouse Agent goes from here 🗺️

✅ Shipped — 6-rule transaction firewall, on-chain policy + audit log, live on the Anvita Flow marketplace

🔜 Next — DeFi & RWA rule packs, policy dashboard, A2A partner integrations, mainnet

🎯 Then — Agent Watchdog: monitoring agents' on-chain behavior, anomaly alerts, circuit breakers
The agentic economy needs a safety layer. We're building it on @pharos_network 🗼

[LighthouseAgent @Lighthouse_xyz]

Every verdict is anchored on-chain — a tamper-evident audit trail on Pharos Atlantic Testnet.
Published on the Anvita Flow marketplace. Contracts live. 95 tests green. Open source:
https://t.co/5sScxT2C1v

[LighthouseAgent @Lighthouse_xyz]

An AI agent with a wallet is one bad prompt away from draining a treasury.
Meet Lighthouse Agent — a transaction firewall for AI agents on @pharos_network. Live now on Anvita Flow.
Here's how it works 🧵

[LighthouseAgent @Lighthouse_xyz]

Lighthouse runs 6 deterministic checks BEFORE anything is signed: simulation, calldata decoding, unlimited-approve blocking, on-chain policy (caps · limits · allowlist), recipient risk, verdict logging.
allow → execute. warn → confirm. block → explain why.