The doctrine · Updated July 2026
The Agent Management Method
Nine principles for running fleets of AI agents the way great organizations have always run people. This is the doctrine Vivari is built on — useful whether or not you ever use Vivari.
01Context before work
Never let an agent start cold. A brilliant stranger interrogating your repository from zero will be confidently wrong about your conventions, your history, and your landmines — every single session. Onboarding is not a nicety; it is the difference between an agent that ships your intent and an agent that ships plausible fiction. Give every agent a briefing it can trust before it touches anything.
02Memory is judgment, not storage
"Remembers everything" is a warehouse, not a colleague. Real memory is retrieval judgment: knowing what to surface, when, triggered by what, at what depth. And memory has classes that must never be pooled: a learned preference should bend under context; a hard gate must never bend; standards and rules sit in between. What one agent learns on Tuesday, the right agent should know on Wednesday — at the right level of detail, and not one word more.
03Permission follows role
Nobody hands the new hire production keys on day one — yet the industry default for agents is everything, everywhere, immediately. Invert it: each agent ships as a package — role, instructions, and a permission profile scoped to its niche. A reviewer that cannot write. A writer that cannot deploy. Scope is not distrust; scope is what makes delegation safe enough to be generous with.
04Evidence over opinions
When an agent's change is risky, the review must produce evidence, not vibes — and an LLM reviewing an LLM produces vibes with confidence. Deterministic checks against your own history ("these files change together in 14 of 15 commits; this diff touches one") give you findings that are reproducible, traceable, and defensible in an audit. Save the model's judgment for the work; let facts judge the risk.
05Review at both ends
Quality control belongs at both ends of the pipeline: curate ideas before they become work, and gate changes before they ship. And upstream curation works best cross-model — one model family drafts, a different one adversarially reviews — because independent priors catch what shared blind spots never will. A good organization doesn't let a junior's first raw idea go straight to production planning; neither should yours.
06Everything on the record
Every tool call, every message, every decision — recorded, replayable, exportable. Not because you distrust the fleet, but because accountability is what lets you trust it more tomorrow than today. An action that isn't on the record didn't happen, and can't be learned from. The audit trail is not compliance theater; it is the fleet's own memory of how it behaves.
07Autonomy is earned, not granted
Start manual. Watch closely. Widen scope as the record justifies it — the same way trust works with people. A clean run of reviewed work is a promotion case; a caught mistake is calibration, not failure. The endpoint is not maximum autonomy — it is correctly allocated autonomy: each agent trusted exactly as far as its record carries, with a human able to take the wheel at any moment, on any agent, without tearing anything down.
08An ecology, not a pyramid
Human org charts are compression artifacts of human constraints — salaries bundle tasks into roles, attention limits spans, titles name the bundles. Agents unbundle. Don't baptize agents as pretend executives; build an ecology of narrow specialists, each shipped with exactly the permissions, tools, and knowledge its niche requires. Supervise by magnification — the whole fleet at a glance, any single agent up close — not by middle management.
09Attention is the budget
Tokens are cheap and getting cheaper; your attention is not. A managed fleet spends both on meters — and the system's job is to earn the right to interrupt you less. The mark of a well-run fleet is not how much it does while you watch, but how little it needs you to watch: every alert that respects your attention builds the trust that lets the next task run unattended.
These principles are old. That is the point — we already know how to absorb brilliant strangers; we've done it with people for a century. The manifesto makes the argument; the definitive guide maps the discipline; Vivari is the place where all nine run as one workspace.
Vivari is the method, running. Curated early-access cohorts open this fall.
Request early access