The factory, checked

Every check here queries the live substrate or the shipped source — none of them greps for a symbol that is defined and never called. Each one reports its own reason, and an unreachable substrate is reported as cannot run, never as a failure.

15 ok · 0 red · 1 cannot run

Receipt generated Wed, 29 Jul 2026 07:13:56 GMT — cloud substrate flsiu1-0

Check State Reason
schema-live the ladder is in the DEPLOYED thing-type @values ok
sdk-enum the generated zod enum accepts the ladder (2nd gate) ok
tracer ONE row written THROUGH the receiver proves schema→query→surface end to end ok
write-path an objective row carries a rung: tag — factory:elaborate wrote it, not a migration ok
seeded vision.md's rows exist as objective things ok
driver ready-tasks() returns at least one real row ok
gaps-visible uncovered() names the dark wires — absence made visible ok
belief-guard the belief functions refuse to gate while their edge is empty cannot run no TYPEDB_TEST_DB scratch database — the evidence wire needs a real substrate to prove, and must never be proven against the production db
fn-exposed the factory funs are in FN_MAP and on the one shared allowlist ok
fn-entity-args fn:run can bind an entity param (the driver funs) ok
fn-one-allowlist web + MCP + CLI + channels read ONE allowlist source, not four copies ok
view-gaps the gaps board queries the ladder AND is honest when empty ok
stream /factory streams what is being built AND can stop it ok
walk-speed the deterministic lifecycle walk passes, with per-stop speed budgets ok
templates text/template-*.md teach the corrected contract; YAML ratchet holds ok
docs plan/vision/docs no longer claim the factory is unarmed ok

A red check failed for a reason you can act on. A cannot run means the substrate was unreachable and nothing was proven either way — it is never a failure, and re-scoping correct work because of one is the mistake this state exists to prevent.

The gaps
5 uncovered
  • Add the missing 0 1 / 0 3 / 0 4 cron triggersdeliverable:vision:2
  • Seed scheduled_jobs rows for learning:dream + learning:evaldeliverable:vision:3
  • Set CAPABILITY_GRAPH=on for the channels workerdeliverable:vision:4
  • Tracer — the one row that proves the chain end to enddeliverable:factory:999
  • Transfer the WorkerCap to WORKER_PRIVATE_KEY_SUI's addressdeliverable:vision:5

Ready to work on

1 tasks
  • Import handleScheduled into the worker entry (astro.config.mjs)task:019fabd296a31ce8e3fbaac3
The run
no reading

Couldn't read the graph

the substrate did not answer — so nothing is known about the run either way. This is not "nothing is running".

No earlier reading is held, so the last known state of the run is unknown.

No reading yet.