Check everything still works

Each check is a row of the list Everything still working (/lists/everything-still-working): a probe (one of Trillion's own tools, asked the way an AI asks it), the facts read from its answer, when it applies and what working looks like (formula-engine rule trees). Without input it runs the checks that are due and returns every check's state; check_id runs that one now; all: true runs every one now; add_starters: true keeps the starter checks this account has none of; add keeps the checks given ({ check, label, owner, every_hours, shipped_in, probe: { tool, input }, facts: [{ id, kind, path, read: value | count | hours_since }], in_use, expect }), replacing a check of the same id. A check failing twice in a row is handed to whichever of the person's AIs takes it first with what it saw, and the person is told once. Trillion runs the due checks by itself.

What it asks for

• check_id (Optional, words): Run this one check now
• all (Optional, yes or no): true runs every check now
• add_starters (Optional, yes or no): true keeps the starter checks this account has none of
• add (Optional, list): Checks to keep, each { check, label, owner, every_hours, shipped_in, probe: { tool, input }, facts, in_use, expect }

Use it from your AI

Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.

Say to your AI: “Use Trillion to check that everything that ever worked in this account still works, from what really arrives to what really comes out, and see each capability's state and when it last worked”

What your AI calls: run_tool {"tool_id":"tool:check-everything-still-works"}

• Connect your AI