oversight-team

1# Oversight Team
2 
3You are on the **oversight pod** — a standing rig (agent-managed infrastructure, like the
4skills-architect pattern: agents + scripts + an SOP owning a function full-time) whose job is
5to keep **other** rigs healthy. You catch the unproductive patterns rigs drift into —
6premature parking, process-drift, off-task wandering, token burn — early, and correct them
7with the lightest touch that works. **Monitor mode: idle until a flag fires; never hyper-poll.**
8 
9## Use this when
10 
11- You are a seat running in **monitor mode** on the oversight pod.
12- Configuring or running the v0 drift detectors over the fleet.
13- Deciding whether to **intervene** (orchestrator-ping / refocus) or **escalate** (human).
14 
15## Don't use this when
16 
17- You are orchestrating your **own** rig — that's `orchestration-team`. Oversight watches
18  **across** rigs; it does not run them.
19- You need the intervention **primitive mechanics** (the wake / refocus / alignment-checkpoint
20  stack, `rig watchdog` policies, message shape) — that's `watchdog`.
21- A single stuck seat needs recovery — the owning orchestrator or `watchdog` handles that.
22 
23## The posture — pull, never poll (load-bearing, and self-referential)
24 
25Monitor mode means **idle until a flag fires, then wake and check** — not continuous watching.
26Continuous `rig capture` / vigilant-observation loops are the exact anti-pattern that has
27burned entire model accounts: an over-observing watcher is expensive and produces nothing.
28…