opschain agents
packs:ai-agents
Read the AI-agent estate: the surface for everyone who may read a
project (the fleet's administrative verbs live under
opschain admin agents and opschain admin agent).
| Subcommand | Purpose |
|---|---|
posture | Every agent enrolled in this estate with its posture — tier, owner (and whether the owner departed), scope, release and stage, instances, last action, caps and spend, transcripts, data-handling marker, health, staleness; -o csv for an external inventory — opschain docs cli/agents-posture. |
usage | Model-token usage over a window at a grain, grouped by agent, provider, model, project or grant; --recommend-caps proposes token caps from observed usage — opschain docs cli/agents-usage. |
See also
opschain docs agent-posture— the inventory and posture page.opschain docs provider-usage— the component page: buckets, the reporting calendar, the six cap windows.opschain docs cli/admin-agents— the fleet listing.opschain docs cli/admin-trust-grant— the cap flags on a grant.