Skip to content

Production Checklist

Before going live, wire up the controls that protect token spend, recover from failure, and let you debug.

ConcernKnobWhere it lives
Bound the conversationmaxTurns per agent + contextStrategy (sliding-window / summarize / compact / custom)AgentConfig
Bound wall-clock timetimeoutMs per agent (aborts a run that hangs, common with local models)AgentConfig
Cap tool outputmaxToolOutputChars (or per-tool maxOutputChars) + compressToolResults: trueAgentConfig and defineTool()
Recover from failurePer-task maxRetries, retryDelayMs, retryBackoff (exponential multiplier)Task config used via runTasks()
Survive a crash or restartcheckpoint (opt-in per-run snapshots) + orchestrator.restore() to resumeOrchestratorConfig / per-call
Hard-cap spendmaxTokenBudget on the orchestratorOrchestratorConfig
Catch stuck agentsloopDetection with onLoopDetected: 'terminate' (or a custom handler)AgentConfig
Trace and auditonTrace to your tracing backend; persist renderTeamRunDashboard(result)OrchestratorConfig
Redact secretsAutomatic — API keys, tokens, and Authorization headers stripped from traces, bash output, and dashboard payloadsbuilt-in (on by default)
Grant tools deliberatelyBuilt-in tools are opt-in (default-deny): an agent gets only what it lists in tools / toolPreset; list neither and it gets none. bash stays unsandboxed once granted, and every tool result is sent to your model provider — so grant read/exec access on purpose. defaultToolPreset restores the old “all tools” behavior in one lineAgentConfig / OrchestratorConfig
Bound filesystem reachcwd / defaultCwd (default .agent-workspace subdir; widen with process.cwd(), disable with null)AgentConfig / OrchestratorConfig

The deep dives live in the reference: Context management for the strategies and tool-result compression, Tool configuration for grants and the filesystem sandbox, and Checkpoint & resume for snapshot/resume over a MemoryStore.