tool use and context

8 articles · 15 co-occurring · 0 contradictions · 0 briefs

Both implementations involve agents using tools (Python sandbox, Slack API) where context must flow across tool call boundaries

Related concepts

tool integration patterns 8 multi agent orchestration 6 context window management 4 state management 2 prompt engineering 2 state preservation across operations 1 state persistence across steps 1 role based prompt engineering 1 retrieval for context 1 rag retrieval failures 1 prompt engineering vs context engineering 1 plugin architecture 1 natural language interfaces 1 multi turn context management 1 modular context architecture 1

Evidence chain (8 articles, showing 8)

@charmaine_klee: I just published 2 fun cookbook for @claudeai Managed Agents! example_of

Both implementations involve agents using tools (Python sandbox, Slack API) where context must flow across tool call boundaries

What Is Context Engineering? The Complete 2026 Guide - Articsledge supports

Reference to 'ignores a tool it was given' points to context architecture problems in how tools are presented, prioritized, and made available to the model.

Useful AI Agent Case Studies: What Actually Works in Production supports

Identifies tool misuse as context failure: agents need right context to invoke tools correctly

Model Context Protocol (MCP): A New Standard for AI Agents | by Gokcer Belgusen | Garanti BBVA Teknoloji | Medium supports

MCP servers are essentially tools that provide context to AI agents. The protocol standardizes how agents discover, invoke, and receive results from context-providing tools.

Claude Code MCP Tools: 10,000+ Tools via One CLI Command | MCP Bundles supports

Demonstrates that effective tool use requires not just tool availability but context-aware tool selection—Claude decides which of 10,000 tools to use based on the request, which requires understanding

Build AI Agents With Langchain, Crewai & Openai: A Practical ... supports

Building agents with these frameworks inherently involves tool integration, which requires clear context about what information tools need and what they return.

@NirDiamantAI: OpenHands resolves GitHub issues by reading issue descriptions, writing code,... example_of

Agent must use GitHub API tools while maintaining context about the issue, code written, and test results—revealing tool integration context complexity

Agentic AI: Build a Multi-Agent Application with CrewAI supports

Multi-agent frameworks typically involve assigning tools to agents and managing context about which tools are available to which agents

query this concept

$ db.articles("tool-use-and-context")

$ db.cooccurrence("tool-use-and-context")

$ db.contradictions("tool-use-and-context")