1046 posts · Page 52 of 88. ← Blog
Stack the guardrails — callbacks, model filters, restricted tools, and clean-room sandboxing — so that if one layer misses, the next one catches
Layered defense-in-depth for agents: input/output guardrails via callbacks, Gemini safety settings, restricting tools, and sandboxing untrusted actions.
A LangGraph node is just a function — it reads the whole state and returns only the channels it changed.
A node is just a function: it receives the whole current state and returns only the channels it changed. Understand the partial-update contract and the immutable-snapshot guarantee that makes supersteps safe.
This lesson builds a Foundry host agent that discovers remote A2A agents by their cards and calls each one as a tool.
A Foundry host agent resolves remote A2A agent cards and turns each remote agent into a callable tool with agenttool.New over the a2aprovider.
Wrap every run in an OpenTelemetry span, gate risky tool actions behind a permission handler, and swap model providers behind one agent.Agent.
Wrap every Microsoft Agent Framework run in an OpenTelemetry span, gate risky tool actions behind a permission handler, and swap Anthropic, OpenAI, Gemini, Copilot, and Azure behind one agent.
"Half the money I spend on advertising is wasted; the trouble is I don't know which half" is a century-old lament that still captures marketing measurement's core problem. Engineers, arriving with a "measure everything" instinct, often assume the answer is just better tracking — and then discover that the most valuable marketing (brand, demand creation, word-of-mouth) is precisely the hardest to measure, while the easiest-to-measure activities aren't always the most valuable. Measuring marketing well means navigating that tension, not pretending it away.
'Half the money I spend on advertising is wasted; I don't know which half' still captures marketing measurement's core problem. Engineers assume the fix is better tracking — then discover the most valuable marketing (brand, demand creation, word-of-mouth) is the hardest to measure. Measuring well means navigating that tension, not pretending it away.
Go treats errors as ordinary values, not exceptions — which means everything you know about passing, comparing, and inspecting values applies. This is a working guide to sentinel errors, wrapping with %w, and the two verbs that make error chains navigable: errors.Is and errors.As.
Go treats errors as ordinary values, not exceptions — which means everything you know about passing, comparing, and inspecting values applies. This is a working guide to sentinel errors, wrapping with %w,...
How a custodian proves it holds what it owes, what the cryptography actually guarantees, and where the guarantee stops.
How a custodian proves it holds customer assets: Merkle tree of liabilities, on-chain reserve attestation, and the limits of PoR.
Modeling limit/market/stop and GTC/IOC/FOK semantics as a lifecycle instead of a pile of flags.
Model order-type and time-in-force semantics (limit/market/stop; GTC/IOC/FOK) as a lifecycle.
Handle a live market-data feed the way an engineer must: sequence checking, gap recovery, and snapshot-plus-delta order-book reconstruction.
Handle a market-data feed: sequence checking, gap recovery, and snapshot-plus-delta order-book reconstruction.
Enforce pre-trade risk limits and keep real-time positions and mark-to-market P&L without stalling the order path.
Enforce pre-trade risk limits and keep real-time positions and mark-to-market P&L.
How OpenTelemetry traces, structured logs, and token metrics turn an agent's event stream into something you can debug in production.
Seeing inside a running agent: OpenTelemetry tracing with spans for agent, model, and tool steps, structured logging, and exporting traces to debug latency and tool-call trajectories.
The one idea that makes everything else in LangGraph click: nodes don't pass messages, they update a shared state — and reducers decide how.
State is a typed dict of channels; each channel has an optional reducer. No reducer overwrites; a reducer (like add_messages or operator.add) combines. This is the single idea the rest of LangGraph is built on.