# lesse-anthropic-building-agentic-systems-claude-2025-11-23

## Veille

Building Powerful Agentic Systems with Claude - Architecture and Implementation Patterns - Anthropic

## Titre Article

Building Powerful Agentic Systems with Claude

## Date

2025-11-23

## URL

https://www.youtube.com/live/cMSprbJ95jg?si=4HnxK8w1ELvSr4tz

## Keywords

AI agents, Claude, Anthropic, agentic architecture, Claude APIs, agent development, orchestration, developer platform

## Authors

Katelyn Lesse

## Ton

**Profile**: Senior engineer at Anthropic, deep technical perspective, expert register, structured presentation for experienced developers

**Style**: Precise and methodical technical presentation. Professional and didactic tone, using concrete implementation examples. Technical but accessible language, with metaphors related to software architecture. Authority comes from direct experience on the Claude platform. Target audience: developers and software architects building complex agentic systems.

## Pense-betes

- **3 pillars to maximize Claude**: Harness capabilities, Manage context window, Give Claude a computer
- **Thinking as API feature**: Performance scales with reasoning time, configurable token budget
- **MCP (Model Context Protocol)**: Standard for interaction with external systems (GitHub, Sentry)
- **Memory tool**: Client-side file system for storing/recalling relevant context
- **Context editing**: Automatic cleanup of obsolete tool results (39% performance gain)
- **Code execution tool**: Secure sandboxed environment on Anthropic servers
- **Agent Skills**: Folders of scripts/instructions with domain expertise (e.g., design system)
- **Infrastructure challenges**: Container orchestration, session persistence, security at scale

## RésuméDe400mots

Katelyn Lesse, engineering lead for the Claude developer platform at Anthropic, presents a three-pillar approach to maximizing the performance of agentic systems with Claude, using Claude Code as a concrete implementation example.

The first pillar consists of **harnessing Claude's capabilities** through customizable API features. Claude has developed a "thinking" capability where its performance improves with the amount of reasoning time allocated. The API exposes this feature with a configurable token budget, allowing developers to choose between fast responses and in-depth reasoning. Reliable tool use allows Claude to call custom tools, as Claude Code does with its many file manipulation and test execution tools.

The second pillar concerns **context management**. Getting the right context at the right time is critical for performance. Three major innovations support this: MCP (Model Context Protocol) enables standardized interaction with external systems such as GitHub or Sentry; the Memory tool offers a client-side file system where Claude intelligently stores information (codebase patterns, git preferences) to recall when needed; Context Editing automatically cleans up obsolete tool results, generating a 39% performance gain in internal benchmarks.

The third pillar, the boldest, advocates **"giving Claude a computer"**. Rather than debating the complexity of agent harnesses, Anthropic provides the infrastructure enabling Claude to write and execute code autonomously. The Code Execution Tool offers a secure sandboxed environment on Anthropic's servers, managing containers and orchestration. This approach enabled the launch of Claude Code on web and mobile, solving complex orchestration and session persistence challenges.

**Agent Skills** enrich this autonomy by providing domain expertise. These are folders of scripts and instructions that Claude can invoke depending on context, such as automatically applying a design system when creating landing pages. This combination of Skills with MCP creates a system where Claude has access to tools, context, and the expertise to use them effectively.

Anthropic's future vision continues to evolve around these three axes: exposing new Claude capabilities via the API, improving context management with more sophisticated tools, and solving infrastructure challenges to enable genuine secure agentic autonomy at scale.

## GrapheDeConnaissance

- Katelyn Lesse —travaille_chez→ Anthropic (ORGANISATION, 0.98)
- Anthropic —a_créé→ Claude (TECHNOLOGIE, 0.98)
- Claude —utilise→ thinking API (TECHNOLOGIE, 0.95)
- MCP —permet→ interaction standardisée avec systèmes externes (CONCEPT, 0.95)
- Context Editing —mesure→ gain de performance de 39% (MESURE, 0.9)
- Claude Code —utilise→ Memory tool (TECHNOLOGIE, 0.9)
- Anthropic —permet→ Code Execution Tool sandboxé (TECHNOLOGIE, 0.92)
- Agent Skills —améliore→ autonomie agents (CONCEPT, 0.88)
- Agent Skills —converge_avec→ MCP (TECHNOLOGIE, 0.85)
- temps de raisonnement —améliore→ Claude (TECHNOLOGIE, 0.92)
- Katelyn Lesse —recommande→ trois piliers systèmes agentiques (CONCEPT, 0.95)

---
Canonical: https://www.thekb.eu/en/fiches/lesse-anthropic-building-agentic-systems-claude-2025-11-23/