watchdog

1# Watchdog
2 
3Continuity checks are timed or evidence-triggered interventions that
4keep an agent workflow from going idle, drifting, or entering needless
5bureaucracy. They operationalize the refocus doctrine.
6 
7## Use this when
8 
9- **Authoring a watchdog policy** for a long-running rig (artifact-pool
10  edges, stale-owner detection, periodic reminders).
11- **Choosing the intervention level** for a stale-owner situation: is
12  this a wake, a refocus, or an alignment checkpoint?
13- **Drafting wake/refocus/alignment-checkpoint message text** that
14  prevents cadence pollution and bureaucracy theater.
15- **Deciding cadence** — scan vs wake intervals, conservative vs
16  aggressive nudging.
17 
18## Don't use this when
19 
20- The agent is actively working and closing artifacts. Prefer no-op.
21- The intervention is masking bad startup or workflow design — fix the
22  underlying cause, don't compensate via watchdog.
23- The work has a clear next-action handoff already in queue. Use queue
24  nudges instead of a watchdog policy.
25 
26## The 3-level intervention stack
27 
28| Level | Goal | When | What it does |
29|---|---|---|---|
30| **Wake** | Restart motion | Owner appears idle, stale, blocked, or missing a next handoff | Small liveness nudge — does NOT reframe the work |
31…