A weekly scorecard over every loop, cron job and heartbeat we run: which ran, what they cost, which are stalled, noisy, or no longer earning their keep — and which should be killed.
Weekly scorecard to Nish. Killing a useless loop is reported as a success, not a failure — that's the doctrine.
The blog-post rule made standing process: review usage, stop agents when needed. And it works — its very first run found two real, pre-existing failures in jobs everyone assumed were fine (one out of quota, one with a broken path). The roadmap deliberately ends at Wave 4; from then on, this loop sources new candidates from observed gaps and the kill-list.
Sunday evening
Cron jobs, heartbeats, state files, cost logs
Against the five verdicts
With kill-candidates named
Loops live or die by human call
For implementers and the technically curious. The full build sheet — verified queries, thresholds, and build notes — lives on the specs page.
| System | Role in this loop |
|---|---|
| Hermesagent runtime on our server | The scheduler that wakes the loop up. Each loop is a cron job under a Hermes profile; the planned bizops profile will host the business digests (IT loops run under vpsops). |
| Healthcheckshealthchecks.huxapps.com | The dead-man's switch. The loop pings it only after a clean run — if the loop dies or errors, the ping stops and Healthchecks raises the alarm independently. This is how 'never silent' is enforced by machinery, not promises. |
| Usage/cost logsper-run token JSON | Per-run cost files emitted by the agent runtime — IT-3 aggregates what no dashboard does. |
| Email renderer + gwsrender_email.py | All digests pass through one shared renderer: Huxberry-branded HTML, tables for repeated rows, a coral 'needs your response' box when the loop has questions, and an arrow link on every record. Sent from the loops mailbox via the Google Workspace CLI. |
| Loop chassisloop_common.py | Shared plumbing every loop reuses instead of reinventing: state files, run-over-run diffing, Metabase drill-down link building, quiet-on-green notify logic. |
| Model / brain | What it does here |
|---|---|
| Claude Opusthe judgment model | Reviews alert wording, thresholds, and anything a human will read and act on. Post-Fable, Opus owns everything that ships. |
| GPT-5.5 via Codexthe bulk-work model | Wrote and maintains the mechanical parts — SQL, diffing, digest assembly. Effectively free on our existing subscription, so routine cycles cost almost nothing. |
Prior scorecards — verdicts trend over weeks, so “noisy” is a pattern, not a bad day.
Weekly scorecard email/Telegram to Nish.
Read-only over every other loop's state. Its recommendations are the input to a human decision, never an action.