[M3 Request] Claude Code 2.1.154+ sends MiniMax-M3 a reduced system prompt by default — please add CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT=0 to your docs

打开 issue 中链接的 Claude Code 文档页面，并找到现有的 MiniMax 设置环境变量代码块。添加值为 0 的 CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT，保留 JSON 格式，并验证示例仍然有效且符合所请求的配置。

Summary

Claude Code has two system prompts: full and lean (short). Since v2.1.154 (released 2026-05-28), the lean prompt is the default for all non-Claude model IDs. So MiniMax-M3 gets the short prompt. The short prompt removes many instructions that teach the model how to work as a coding agent.

Result: MiniMax-M3 looks much weaker inside Claude Code than it really is. One environment variable fixes this. It is missing from your docs (https://platform.minimax.io/docs/token-plan/claude-code).

How to verify (5 minutes)

1. Install Claude Code >= 2.1.154 and connect it to MiniMax as described in your docs.
2. Run claude in any folder. Type /context. Note the "System prompt" size: about 1.7k tokens. This is the lean prompt.
3. Add this key inside the "env" block of ~/.claude/settings.json. Do not replace the file — only add this one key:

"CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT": "0"

Example of a full file (based on your docs):

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "YOUR_MINIMAX_API_KEY",
    "ANTHROPIC_BASE_URL": "https://api.minimax.io/anthropic",
    "CLAUDE_CODE_AUTO_COMPACT_WINDOW": "1000000",
    "CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT": "0"
  }
}

4. Restart claude. Run /context again: about 7k tokens. This is the full prompt.
5. Optional exact proof: dump raw API requests with OTEL_LOG_RAW_API_BODIES=file:/tmp/ccdump and compare the system field between the two runs.

Measurements

Two live runs, clean config (no user MCP servers), empty folder, Claude Code 2.1.251, model MiniMax-M3. Verified with raw request dumps, 2026-08-31.

Note: exact totals can shift slightly when Claude Code enables or disables a server-side experiment tool, but the lean/full split stays the same.

What MiniMax-M3 loses in lean mode

• The "Doing tasks" section: the engineering discipline rules (prefer editing existing files, no abstractions beyond what the task requires, no half-finished work, test UI changes in a browser, and more).
• Prompt injection detection for tool results (WebFetch / MCP / web content). Completely absent in lean mode (verified in the dumps: 3 mentions in the full prompt, 0 in the lean prompt).
• Rules for parallel tool calls (reduced from a detailed section to a single line).
• Delegation guidance from the system prompt — when to use subagents and the Explore agent (the shortened Agent tool description keeps only the bare core).
• Most of the Bash tool description: git workflow, PR creation steps, background task handling (-87% of characters).

Anthropic wrote in the v2.1.154 release notes: "The lean system prompt is now the default for all models except Haiku, Sonnet, and Opus 4.7 and earlier."

This works for their newest Claude models, which are trained to follow these rules without the explicit text. For MiniMax-M3 the result is different: in everyday use, the same model follows instructions much better with the full prompt — fewer mistakes and fewer repeated steps. Repeated steps also consume Token Plan quota faster.

Why the settings in your docs do not fix this

Your Claude Code page recommends ANTHROPIC_BASE_URL, ANTHROPIC_AUTH_TOKEN and CLAUDE_CODE_AUTO_COMPACT_WINDOW. None of these change the prompt shape. We also verified with raw request dumps that CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 does not restore the full prompt: the CLI applies the lean prompt to non-Claude model IDs unconditionally.

Only CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT=0 restores the full prompt. This is an official, documented option. From https://code.claude.com/docs/en/env-vars: "Set to 0, false, no, or off to opt out even on models where the experiment or server configuration would otherwise enable it. The full tool set, hooks, MCP servers, and CLAUDE.md discovery remain enabled."

Request

Please add "CLAUDE_CODE_SIMPLE_SYSTEM_PROMPT": "0" to the env block in https://platform.minimax.io/docs/token-plan/claude-code (and to any MiniMax coding-plan setup tooling, if one exists).

One line, officially supported by Claude Code, and it restores full agent quality for all Token Plan users in Claude Code.

References

• Release notes v2.1.154 (published 2026-05-28): https://github.com/anthropics/claude-code/releases/tag/v2.1.154
• Official env vars docs: https://code.claude.com/docs/en/env-vars
• Independent comparison and security note (same split measured on a Claude model): https://github.com/CherryHQ/cherry-studio/pull/18028