Blog
Writing
Life, tech, and everything in-between.
Guardrails for LLM Agentic Systems: Layered Defense
Safety instructions in a system prompt are not sufficient for production. Benchmark agents still behave unsafely in a significant fraction of cases. The fix is layered defense: prompt-level, tool-level, model-level, and system-level guardrails stacked so any one failure is caught by the next.
Human-in-the-Loop Patterns for LLM Agents
Full autonomy is not the right target for every agent. Put a human in the loop at the points that matter: before dangerous actions, for ambiguous decisions, as an escalation path when the agent is stuck. Checkpointing makes the pauses cheap.
Skills as Contextual Memory: Reusable Procedural Knowledge for LLM Agents
Anthropic's Skills pattern packages reusable procedural knowledge as folders of instructions, scripts, and resources. Progressive disclosure loads only the metadata at startup and brings in the details on demand. Context capacity becomes effectively unbounded.
Persistence and Checkpointing: Time Travel and Recovery for LLM Agents
A long-running agent that loses its state on the next deploy is not a production system. Checkpointing saves agent state after every step, enabling conversational memory, human-in-the-loop pauses, time travel for debugging, and fault-tolerant resumption.
Generative Agents Memory: The Stanford Architecture for Persistent LLM Agents
Stanford's Smallville study ran 25 LLM agents that formed relationships and organized events autonomously. The mechanism that made their behavior coherent over days was a three-part memory system: observation, scored retrieval, and periodic reflection.
Short-Term and Long-Term Memory for LLM Agents
An agent with no memory repeats itself. An agent with bad memory contradicts itself. Getting memory right means splitting it into kinds: short-term for the session, long-term for durable facts, procedural for how-to, episodic for what-happened.
Group Chat Patterns: Round-Robin, Selector, and Swarm for LLM Agents
When several agents share a conversation, the question becomes: who speaks next? Microsoft AutoGen names four ways to decide. Each produces a different coordination style, with different strengths on different workloads.
The Shared Scratchpad: A Common Workspace for Collaborating Agents
When agents need to see what each other are doing, neither a supervisor nor a handoff will do. Give them a shared workspace instead. Every agent reads from and writes to the same state. Context sharing is maximal, but so is the risk of distraction.
Handoffs and the Swarm Pattern: Peer-to-Peer Agent Transfer
A supervisor is not the only way to coordinate agents. In the swarm pattern, agents transfer control to each other through handoff tool calls. No central manager; the conversation simply moves to whichever specialist it needs to be with.
Hierarchical Teams: Supervisors of Supervisors in Multi-Agent Systems
A supervisor with five specialists works. A supervisor with twenty does not. Hierarchical teams stack supervisors, each responsible for a department of specialists, and keep a dual ledger to track plan and progress separately.
Supervisor and Router: A Central Agent That Delegates to Specialists
When a single agent carries too many tools and too many responsibilities, the fix is not a bigger prompt but a supervisor whose tools are other agents. Each specialist has its own prompt, its own tools, and its own scratchpad.
LATS: Monte Carlo Tree Search for LLM Agent Decisions
LATS brings Monte Carlo Tree Search, the algorithm behind AlphaGo, to language agents. Reasoning, acting, and planning become a single tree that the agent navigates by balancing exploration against exploitation. Accuracy climbs, cost climbs with it.