$BORT The part of BORT I like most is the custody design. Putting funds in each agent’s ERC-6551 wallet, with capped, expiring, revocable spend grants, means the worst case is bounded by what the owner granted. My question is about the layer those caps don’t cover. Hunter agents read signals from public Telegram groups and X, then trade on them, a pump group or a coordinated account can post a “signal” written to push a model into buying a token. So what sits between a scraped signal and an executed trade?

Is there a minimum liquidity or holder-distribution check before a buy?

Is there a per-token position cap that’s separate from the overall spend cap?

Can an owner require manual approval above a set size?

The docs say a third-party audit hasn’t been published yet. I appreciate that honesty. When it happens, I’d like the scope to cover the runtime and signal-handling logic, since that’s where most of the risk is.