Changelog

v0.2.0

2026-03-21

createCostTracker:Stateful per-model cost tracker with O(1) budget checks, cost snapshots, and reset. Replaces the stateless budgetExceeded for long-running agents.
DEFAULT_MODEL_PRICING:Built-in pricing map for 13 models: OpenAI (gpt-4o, gpt-4.1, o3, o4-mini), Anthropic (claude-sonnet-4, claude-3-5-haiku), Google (gemini-2.5-pro/flash, gemini-2.0-flash).
extractPendingApprovals:Extract tool calls awaiting human approval from generateText results.
createInterruptHandle:Create serializable JSON handles for interrupted agent state. Safe for KV/DB storage.
resolveInterrupt:Pure function to apply human decisions (approve/deny/edit args) and produce resumption messages.
InMemoryStore:Zero-dependency MemoryStore implementation with lazy TTL evaluation.
EmployeeAgent:Agent class composing membrane + memory + prepareStep + stop conditions. Delegates to ToolLoopAgent.
createAuditLogger:Callbacks for tool call and step observability pipelines.

budgetExceeded now supports per-model pricing via pricing parameter (in addition to flat-rate mode from v0.1).
composePrepareStep merge rules documented and tested: system concatenates, activeTools intersects, model/toolChoice last-writer-wins, experimental_context/providerOptions deep-merge.

2026-03-14

membrane:4-tier permission system (auto, draft, confirm, block) for AI SDK tools.
resolveTier / explainTier:Tier resolution with explicit tiers, custom resolve function, glob patterns, and configurable default.
composePrepareStep:Compose multiple prepareStep functions with deterministic merge rules.
createMemoryPrepareStep:Frozen-snapshot memory injection at step 0 from any MemoryStore.
createHeartbeat:Always-on tick function with concurrency guard and circuit breaker.
budgetExceeded:Stateless stop condition for token and flat-rate USD budgets.
tokenVelocityExceeded:Stop condition for detecting runaway token velocity.