Resources
Resources for Claude Code teams.
Current writing, production guides, and defined vocabulary for teams turning Claude into an engineering practice.
Current thinking
Start with the work that is moving now.
Field notes, implementation lessons, and strategy calls from current Claude work.
Browse by workstream
Production guides
Evergreen references for running Claude Code in production.
Start with the operating model, then move into MCP, hooks, skills, API work, reliability, and governance.
Start here
Browse all guides
Running Claude Code as a Production Engineering Practice
How a consultancy delivers billable client engineering with Claude Code, gated by deterministic validators and demonstrated through published receipts.
Read the production practice guide MCP Servers in Production: Ownership Inversion, the Protocol Vocabulary Limit, and What the Spec Leaves to You An MCP server moves the integration to a system into the server and out of the agent. The inversion pays inside a finite protocol vocabulary; outside it, the protocol is overhead. The spec also draws a second boundary worth naming, where functional integration moves but governance does not. Claude Code Hooks in Production: The Gate the Model Doesn't Get to Skip Why a hook is the layer of an agentic coding stack the model cannot choose to skip, why a fail-open hook can be worse than no hook, and where enforcement belongs. Claude Code Skills in Production: Two-Axed Discoverability and the Patterns That Make Skills Compound A Claude Code skill earns its value through two-axed discoverability, the same module answering a slash command AND a natural-language request. Designing for both surfaces deliberately is the work; catalog hygiene is the pattern that lets skills compound past the threshold where description-routing accuracy starts to decay. Claude API in Production: A Runtime, Not a String Function, and What It Leaves to You The Claude API is a structured runtime, not a text-in/text-out endpoint. It hands you capability primitives but leaves conversation state, cost, and verification of non-deterministic output to you. Owning that boundary is the production work. Agent Reliability in Production: A Verification Loop, Not a One-Time Test Why you cannot unit-test a non-deterministic agent, why reliability is the inverse of your known failure modes, and how the verification loop has to keep re-running, in parallel, as the system evolves. Agentic AI Governance in Production: Who Owns the Bar When the Agent Ships A written policy sets the rules but cannot enforce them on an autonomous agent. Agentic AI governance is the deny-by-default gates, scoped permissions, provenance records, and named verification-bar owners that put those rules in the execution path, between an agent and a shipped artifact.
Glossary
Defined vocabulary for agentic development.
Each definition has its own page, citations, related terms, and guide appearances.
MCP server A program that exposes tools, resources, and prompt templates to Claude Code and other Model Context Protocol clients over a standardized protocol, so an agent can read from and write to external systems without bespoke per-integration code. Claude Code hook A user-defined command or prompt that Claude Code runs automatically at a named lifecycle event, such as before or after a tool call, enabling deterministic enforcement of rules the agent cannot skip. Claude Code skills Versioned capability modules that extend Claude Code with domain procedures, loaded on demand rather than held in the base prompt. Subagent orchestration Decomposing a development task across multiple Claude Code subagents that run independently under a coordinating orchestrator, with explicit context isolation and result aggregation. Verification loop A verification loop is the practice of treating an AI agent's output as unproven until it passes a repeatable verification step, then feeding any failure back into another attempt, so reliability comes from an iterated check-and-correct cycle rather than a single one-time review. Structured outputs Structured outputs are a model capability that constrains a response, or a tool call's arguments, to a supplied schema, so the caller receives validated typed data instead of free-form text it has to parse defensively. Agent evaluation Agent evaluation is the practice of measuring whether an AI agent succeeds at a task by scoring its behavior, its tool calls, and its final outcome against explicit criteria, run as a repeatable harness rather than judged on a one-time impression. Context engineering Context engineering is the practice of deliberately curating what an AI agent holds in its context window, the system prompt, retrieved data, tool results, memory, and conversation history, as an engineered system rather than ad-hoc prompt wording. Tool use A Claude API capability in which the model, given a set of tool definitions, emits a structured request to invoke one and resumes once the caller returns the result, letting an agent act on external systems rather than only generating text. Prompt caching A Claude API feature that stores a repeated context prefix server-side so later requests reuse it at reduced cost and latency, lowering the price of repetitive agentic workflows. Agent memory A persistent store, separate from the context window, that an agent loads at the start of a session and can add to as it works, so facts and preferences survive across sessions instead of resetting when the context window does. Deterministic validator A deterministic validator is a rule-based, pass-or-fail check that enforces an invariant the same way on every run, independent of model judgment, serving as the non-probabilistic floor under an agentic pipeline. Model Context Protocol An open standard that defines how AI applications connect to external tools, data, and prompt templates through one uniform client-server protocol, so any compliant client can use any compliant server without integration code written for that specific pairing. Context window The bounded span of tokens a model can read and reason over in a single request, holding the system prompt, prior turns, supplied material, the current input, and the reasoning and output the model produces. Claude Code Claude Code is Anthropic's agentic coding tool, a terminal, IDE, and desktop surface where Claude reads a codebase, edits files, and runs commands, and that I configure with skills, hooks, subagents, and project memory to fit a production engineering workflow.