Ning Xia·1d agoWhere Knowledge Goes in Your Agent HarnessGive each kind of knowledge its place in the harness before the task starts, so the model’s best guess is the one you expect
Ning Xia·1d agoYour Agent Harness Runs on KnowledgePrior knowledge has two parts: initial knowledge drives the task, and the acquired knowledge policy decides what happens to what the agent…
InModels to SystemsbyDeepak Mishra·Oct 5Wiring Shared Memory Into a Multi-Agent Orchestration Platform: Stop Re-Deriving What You Already…The orchestrator already has specialists, a critic, and a task graph. What it’s missing is a past.
InModels to SystemsbyDeepak Mishra·Oct 5Designing a Multi-Agent Orchestration Platform: Coordination Without Chaos
InModels to SystemsbyDeepak Mishra·Oct 4Designing an Agentic Memory System: Giving LLM Agents a Past Without Losing the PlotHow to let an agent remember what matters, forget what doesn’t, and never let its own memories be weaponized against it.A response icon2
