---
name: auto-maintenance
description: Auto Office 3.0 reference spoke, not loaded by 3.1 runs (they use the office CLI). Internal Auto Office v3 maintenance spoke. Use at the beginning or closeout of an Auto Office invocation to label eligible historical outcomes lazily, refresh stale model catalogs outside route-time, update local harness bindings, maintain hot/warm/dreamt evidence tiers, compute maturity/decay, and prepare replay inputs without requiring an always-running daemon.
---

# Auto Maintenance

> **Auto Office 3.1:** this is 3.0 reference material. A 3.1 run is driven by the `office` CLI and runtime-delivered
> role briefs; do not run the `office_runtime.py` helpers below for it. Follow `office status` and its `next:` line.

Maintenance is event-driven by future Auto Office invocations; no daemon is required.

- Mark an observation eligible at the earlier of 14 days after merge or the next three relevant repo/surface runs, then label on the next maintenance pass.
- Preserve negative-outcome asymmetry: recurrence/revert/material post-merge defects are stronger ground truth than no-observed-failure is positive.
- Maintain hot/warm/dreamt authority by lineage. Dreamt rows keep raw local storage but lose numeric routing authority.
- Trigger catalog refresh when local harness fingerprint changes or catalog snapshot is older than 72 hours. Refresh may use network; current routing does not. If refresh fails, keep the last valid snapshot and record staleness.
- Apply ecosystem decay only to affected evidence lineage.
- Compute maturity deterministically from policy + run database; failures move age backward.
- Do not refit routing policy before the replay/calibration evidence bars are met.