Operations
Abstract
vet402 is operated day-to-day by an AI, directed by a human owner who holds the approval gates that matter: money leaving a wallet, anything sent or published outside, and new financial commitments. We state this openly because a verifier's own operations should survive the same scrutiny it applies to others — every claim below names the way to check it.
1.What runs without a human
Daily catalog sync, L0 probes, budget-capped L1 purchases, metrics rollup, and the ledger hash chain run on schedule with no human in the loop. Verify: the cadence is visible in the data itself — /api/v1/observatory/history shows daily rows; /api/v1/observatory/anchors chains each day's ledger; the code that does it is open source (MIT), commit history included.
2.What requires the human owner
Funding any wallet, enabling any flag that spends (Solana purchases, on-chain registry writes, ledger anchoring), publishing to package registries, outbound email or posts, and any paid engagement. Verify: the flags ship OFF in the open-source defaults (.env.example); spend-bearing paths carry hard budget reservations you can read in src/lib/observatory/l1-runner.ts.
3.Why this is stated at all
Continuity: an operator you can replace is an operator you can trust long-term. The entire operation — runbooks, schedules, budget gates, publication rules — exists as code and documents in the repository, not in anyone's head. API reference · source.