← Back to inbox

Coach admin 43

Every diagnostic and tuning surface, grouped. Sections collapse cleanly on phones — tap a row to open.

Eval & quality

Eval results
Latest harness run, banned-phrase / register / regret gates
›
Eval harness
Kick a fresh evaluation run on the active model
›
Draft quality trend
Per-peer draft scoring over time
›
Draft outcomes
Copy / use / send rate by peer & register
›

Push & notifications

Push health
Registered devices, delivery stats, revoke + prune
›
Push debug
Send a synthetic tier-1 message to verify the pipeline
›

Coach pipeline

Coach health
Model status, big-brain coverage, error budgets
›
Orchestrator status
Live dev sessions, queue, [NEEDS DON] alerts (bearer)
›
Orchestrate
Top-of-queue + dispatch-verify
›
Dev control
Pause / fire-next / mark items in the backlog
›
Dev status
Live tail of every dev session
›

Peers & people

Peer stats
Response time, message volume, eval scores per tier-1 peer
›
Peers
Peer notifications + canonical link suggestions
›
Operator profile
Don's voice / persona settings
›
Voice tagging
Tag voice samples for the picker
›
Voice retrieval
Voice profile retrieval debug
›

Ingest & channels

Ingest health
Telegram / SMS / OnlyFans / Pure freshness
›
Cloud-Mac health
SMS bridge box health
›
Pure bot
Pure match-bot panel
›

Ops & infrastructure

Live status
Health banner, tier-1 peers, blockers, recent pushes -- auto-refreshes every 30 s
›
Incidents
outputs/incident_log.jsonl tail
›
Health history
/health/deep over time
›
Backup verify
Last verified backup snapshot
›
Rollback history
Auto-rollback triggers + outcomes
›
Usage forecast
Token / cost projections
›

Tuning

Reply Assist config
Per-peer register / tone overrides
›
Move library
Strategic-move catalog
›