anthropic-platform-features

# Anthropic Platform Features — Router | Field | Value | |---|---| | **Source docs** | [platform.claude.com/docs](https://platform.claude.com/docs) | > **This skill is auto-updated daily.** A pipeline reads the upstream > docs and rewrites the per-surface files below. Section structure is > stable; content drifts to track upstream. ## When to use Router skill for Anthropic platform-side features that sit above the raw Messages API: Agent Skills format (the spec for `.skill` packages used by platform-managed agents), MCP connectors hosted by Anthropic, tool use (computer use, code execution, bash tool, define-your-own tools), build-with-claude features (extended thinking, message batches, citations, Amazon Bedrock, Google Vertex, embeddings, fast mode, context editing), manage-claude operations (workload identity federation, billing, identity), and the Managed Agents product. Use when the user asks about: writing an Agent Skill `.skill` package, Anthropic's hosted MCP connector, MCP Tunnels (connecting private-network MCP servers without opening inbound ports), the computer-use tool, the code_execution tool, bash tool, extended thinking, message batches, prompt caching strategy at the platform level, citations, deploying via Bedrock or Vertex, embeddings models, fast-mode behavior, context editing strategies, WIF setup, billing / cost reports, identity & SSO setup, customer-managed encryption keys (CMEK via AWS KMS / Google Cloud KMS / Azure Key Vault), the Managed Agents product, or self-hosted sandboxes (running agent sessions in your own infrastructure). Skip: raw POST /v1/messages requests (use anthropic-api), Claude Code CLI (use claude-code), Claude Agent SDK (use claude-agent-sdk), user-facing Claude Connectors directory (use claude-connectors). ## Dispatch table | Surface file | Read when the user asks about… | |---|---| | [`SKILL-agents-and-tools.md`](SKILL-agents-and-tools.md) | Agent Skills format spec, MCP connector, MCP Tunnels (private-network MCP servers), remote MCP servers, tool use (computer use, code execution, bash, define-your-own tools), tool_choice | | [`SKILL-build-with-claude.md`](SKILL-build-with-claude.md) | extended thinking, message batches, prompt caching, citations, Amazon Bedrock, Google Vertex, embeddings, fast mode, context editing, vision; also platform foundation (`intro`, `get-started`) and guardrails / streaming-refusals from `test-and-evaluate/` | | [`SKILL-manage-claude.md`](SKILL-manage-claude.md) | workload identity federation (WIF), billing & usage, organizations & workspaces operations, identity & SSO, customer-managed encryption keys (CMEK) | | [`SKILL-managed-agents.md`](SKILL-managed-agents.md) | Managed Agents product — agent definitions, deployment, lifecycle, monitoring, self-hosted sandboxes | ## Examples <example> Context: A developer wants Claude to run code on the server side. user: "How do I enable the code execution tool, and what does the sandbox allow?" assistant: Uses anthropic-platform-features and reads the agents-and-tools surface for the code execution tool. </example> <example> Context: A developer asks about raw request fields. user: "What is the max_tokens limit on /v1/messages for Opus?" assistant: Routes to anthropic-api instead; raw request fields and model limits live there. </example> --- *This skill is auto-updated daily by a maintainer-run pipeline. File issues at [xiaolai/anthropic-docs](https://github.com/xiaolai/anthropic-docs).*