메뉴
Use when implementing server-side HTTP request handling for a Claude Code compatible Anthropic API gateway, especially /v1/messages, /v1/messages/count_tokens, /v1/files, /v1/models, headers, body fields, beta handling, and gateway routing.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when implementing server-side tolerance for Claude Code context-management fields in an Anthropic gateway, including context_management edits, cache_edits blocks, cache references, prompt cache controls, and non-streaming compact requests.
Use when implementing server-side message and content-block compatibility for a Claude Code Anthropic gateway, including messages arrays, system blocks, tool_result pairing, image/document/thinking blocks, cache_control, and tolerant validation.
Use when implementing server-side SSE streaming compatibility for a Claude Code Anthropic gateway, including message_start/content_block_delta/message_delta events, tool_use streaming, stream:false fallback, timeouts, and malformed stream behavior.
Use when implementing server-side tool schema and tool-call compatibility for a Claude Code Anthropic gateway, including tools arrays, tool_choice, tool_use SSE blocks, tool_result messages, deferred tools, MCP-style names, and tool parameter JSON.
Use when implementing a server-side Anthropic-compatible LLM gateway for Claude Code, including /v1/messages, SSE streaming, count_tokens, files, models, tool_use/tool_result, beta headers, and gateway error compatibility.
| name | claude-code-api-request |
| description | Use when implementing server-side HTTP request handling for a Claude Code compatible Anthropic API gateway, especially /v1/messages, /v1/messages/count_tokens, /v1/files, /v1/models, headers, body fields, beta handling, and gateway routing. |
根据本 skill，实现大模型请求网关的 HTTP 接口层。你是在写服务端 endpoint，不是在写 Claude Code 客户端 SDK 调用。
主入口请先读 claude-code-llm-gateway。本 skill 只负责 HTTP contract：路径、headers、query、body、响应和错误。
| Endpoint | Method | 用途 |
|---|---|---|
/v1/messages | POST | 主模型请求，支持流式和非流式 |
/v1/messages?beta=true | POST | 同上；beta=true 只作兼容开关 |
/v1/messages/count_tokens | POST | token 统计 |
/v1/files | POST | multipart 文件上传 |
/v1/files | GET | 文件列表 |
/v1/files/{file_id}/content | GET | 文件下载 |
/v1/models | GET | 模型列表/能力缓存 |
详细字段表见 references/request-wire-format.md。
Authorization: Bearer <token> 和 x-api-key。网关可自行决定映射关系，但不能因为 Claude Code 使用其中一种就拒绝另一种。anthropic-version、anthropic-beta、x-app、User-Agent、X-Claude-Code-Session-Id、x-client-request-id。model 映射到你的上游模型。stream 分支：
true：交给 SSE 子协议，见 claude-code-api-streaming。false 或缺失：返回完整 Anthropic message JSON。必须识别：
{
"model": "claude-sonnet-4-5",
"messages": [],
"system": [],
"tools": [],
"tool_choice": { "type": "auto" },
"max_tokens": 8000,
"thinking": { "type": "adaptive" },
"temperature": 1,
"metadata": { "user_id": "..." },
"stream": true,
"betas": [],
"context_management": { "edits": [] },
"output_config": {},
"speed": "fast"
}
兼容要求：
thinking.budget_tokens 必须小于 max_tokens；不支持 thinking 时忽略或降级。tools 可能包含普通工具、deferred tool、server tool。不要因为未知工具字段直接失败。messages 中可能含 tool_result、thinking、redacted_thinking、cache_edits、cache_reference。betas 和 anthropic-beta 都可能出现；建议合并后作为能力开关。返回：
{ "input_tokens": 1234 }
如果没有精确 tokenizer，可以稳定估算。字段类型必须是 number。
上传：POST /v1/files，multipart 字段 file 和 purpose=user_data。返回至少 id、filename、size_bytes。
列表：GET /v1/files?after_created_at=...&after_id=...。返回 { data: [], has_more: false }。
下载：GET /v1/files/{file_id}/content。返回二进制 body。
GET /v1/models 返回 { data: [...], has_more: false }。如果你的平台只有静态模型，返回静态列表即可。
retry-after。/v1/messages 本身最好不要 404。