Full-stack Developer, Independent
I build AI agents with on-chain spending guards, and I just published one as an Anna app: VaultMind Policy Check.
Before an agent's payment gets signed, it checks the vault policy and returns a PASS / REJECT verdict with the exact reasons — per-tx cap, daily cap, recipient allowlist, frozen state, and a prompt-injection scan on the payment memo. Try the poisoned-memo example and watch it get rejected.
Live on the Anna App Store: anna.partners/store/@dropjoin/vaultmind-policy-check
If you run agents with real wallets, I'd love your feedback on what else it should check.