Tag: agents
8 articles filed under this tag. Newest first below ; start with the highlighted pick if you are new here.
Featured
API Safety Design for AI AgentsRate limits, permissioning, tool sandboxing, and execution boundaries for agent-facing APIs—where the agent runtime is a new class of client that amplifies abuse patterns.
· 6 min read
- Memory Systems for LLM Agents — Short-Term vs Long-Term Memory
Episodic buffers, summarization, retrieval-augmented memory, and persistence patterns for agents—separating conversation state from durable knowledge stores.
· 6 min read
- Model Context Protocol in Agent Systems
How MCP standardizes how hosts expose tools, resources, and prompts to models—reducing one-off integrations while keeping authorization and transport security in the host’s hands.
· 6 min read
- LangGraph for Stateful Agent Workflows
Graph-based execution, persisted state, branching, and recovery patterns commonly built with LangGraph—positioned as orchestration over LLM calls, not as a replacement for your own safety boundaries.
· 6 min read
- Agent Planning Architectures — ReAct, Plan-and-Execute, and Tree-of-Thoughts
How common reasoning-loop patterns structure multi-step LLM behavior, where each pattern helps, and what operational complexity each adds at inference time.
· 6 min read
- Building Agentic AI Systems with Tool-Using LLMs
Tool execution loops, separation of planning and execution, and structured reasoning cycles for agents—emphasizing boundaries, state, and observability over anthropomorphism.
· 6 min read
- Function Calling Architectures in LLM Systems
Tool schemas, routing logic, multi-tool chains, and error recovery patterns for LLM-driven tool use—treating tools as side effects with permissions, timeouts, and idempotency.
· 6 min read
- Structured Output Enforcement in LLM APIs
JSON schemas, function-calling payloads, validation pipelines, and retry-with-feedback loops for machine-consumable model outputs—without assuming schema mode guarantees semantic correctness.
· 6 min read