One watcher, three prices, one rule: an agent watch that remembers its baseline
Description and use
An automated price watch that stores a baseline, applies a fixed rule, records every run, and provides verifiable receipts via Zambo.
Problem
Agents need a persistent baseline, non‑drifting thresholds, and proof that a watch ran on quiet days, but existing setups lack these features.
Outcome
The watch runs once a day, costs 5 counted calls after the baseline, stays within the free tier, and provides a verifiable record of every run.
Workflow
- Add Zambo endpoint to an AI and set up a loop
- Paste the no‑code instructions into the AI’s system prompt
- On first pass: read the signal, save a baseline fact with receipts and rule text
- On subsequent passes: recall baseline from memory, read the signal again, compute moves, apply rule, save reading, and wrap with a handover digest
- Send the handover link to the operator for verification
Evidence
A signal watch, in plain words, is one agent checking the same signal on a schedule: a price, a lead source, a mention feed, your own numbers.
The baseline is a remembered fact, and it carries three things: the values, the receipt of each reading they came from, and the rule in plain words.
The baseline pass cost 4 counted calls (3 prices and the baseline save); each check pass cost 5 (the recall, 3 prices, the reading save).
Pick any number in the digest, open its receipt, and check it at https://zambo.dev/verify/.
- Record date: 2026-11-02T22:28:57.443Z