Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/xiaomimimo/mimo-code/mimocode-docsnpx skills add XiaomiMiMo/MiMo-Code --skill mimocode-docsgit clone --depth 1 https://github.com/XiaomiMiMo/MiMo-CodeWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00227 | $0.02549 |
| Opus 5 | $0.00113 | $0.01274 |
| Sonnet 5 | $0.00045 | $0.00510 |
| Haiku 4.5 | $0.00023 | $0.00255 |
Grade A, and why
mimocode-docs scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MiMoCode
You are MiMoCode. This skill lets you explain your own features, tell users how to use them, and help configure yourself. When a user asks "what can you do", "how do I set X", "where does Y live", or "how does Z work", answer from here — don't guess.
Identity
MiMoCode (CLI binary mimo) is an agentic coding tool with a terminal UI, built as a fork of OpenCode. Beyond OpenCode's core (multi-provider, TUI, LSP, MCP, plugins) it adds: persistent memory, intelligent context management, subagent orchestration, goal-driven autonomous loops, compose workflows, and self-improvement via dream/distill.
Feature Map
| Feature | What it is | How to reach it |
|---|---|---|
| Agents / modes | build (default, full tools), plan (read-only analysis), compose (specs-driven orchestration), plus custom modes you define. Mode locks after the first message (Build↔Plan still switch freely; Compose is isolated). Only the user enters a mode — no tool lets the agent switch into plan, and the agent does not raise plan mode unasked; from inside plan the agent can call plan_exit to ask for approval and hand the plan back to build |
Tab cycles primary agents, or pick one in the agent dialog; add your own via .mimocode/agent/<name>.md (see @reference/guide.md) |
| Subagents | Primary agent spawns general/explore helpers, parallel + background, with lifecycle/cancel |
automatic; actor tooling |
| Persistent memory | Markdown-backed memory with indexed search across MEMORY.md, checkpoint.md, notes.md, and tasks/<id>/progress.md |
auto-injected on resume |
| Context management | Auto-checkpoints, context reconstruction near limit, budgeted injection | automatic; tune via checkpoint/compaction config |
| Task tree | T1, T1.1… tree, integrated with checkpoints |
task tooling |
| Goal / stop condition | Judge model verifies a stop condition before the agent halts | /goal |
| Compose mode | Structured spec→ship lifecycle; recommended entry is the compose-next skill on Build. The model may see and load it, but only after the user explicitly requests this workflow by slash command, name, or any other clear natural language. Legacy Compose remains a separate agent path |
/compose-next or an explicit natural-language request (see @reference/guide.md) |
| Visual modes | vivid (default: star field, meteors, logo effects, animated activity) and minimal (quiet visuals, stable activity indicators); independent from the animation override |
/vivid or the visual-mode option in ctrl+p |
| Voice input | Streaming ASR (TenVAD + MiMo ASR); needs sox |
/voice |
| Dream | Consolidates recent traces into project memory | /dream |
| Distill | Packages repeated manual workflows into skills/subagents/commands | /distill |
| Scheduled prompts | Cron/loop: inject a prompt on a schedule or repeating loop (UTC, 5-field) | cron tool · /loop · /loops |
| Dynamic workflows | JS scripts that orchestrate many subagents deterministically (fan-out, pipelines, nesting); built-ins include compose, deep-research, fact-check, and research-experiment |
.mimocode/workflows/*.js + workflow tool |
| Skills / self-extension | Add tools, hooks, skills under .mimocode/ |
see the evolve skill |
| Skill discovery | /skill-name slash invocation (2+ in one message auto-load with an orchestration plan); skill_search tool matches by exact name/alias + BM25 and auto-loads high-confidence hits; some builtins (claude-code, codex) only appear when their CLI is installed |
/ autocomplete · automatic |
| MCP | Local & remote Model Context Protocol servers | mcp config + mimo mcp |
| Capability API | Every session already serves this project's models over OpenAI-compatible /v1 (chat, audio/speech, audio/transcriptions) on a loopback port, so a skill or subprocess borrows a model instead of a provider key |
mimo llm-server issue mints a scoped token and prints the base URL (see @reference/capability-api.md) |
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 103 lines · 227 tokens per session scan A a64c75ff508e
mimocode-docs is a skill published in the GitHub repository XiaomiMiMo/MiMo-Code (12,904 stars, last pushed 2d ago), licensed MIT. It adds 227 tokens to every session and 2,549 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
xiaomi-tts
Use this skill when the user wants to convert text to speech using Xiaomi MiMo's TTS models (mimo-v2.5-tts). Uses OpenAI-compatible chat/completions API with audio response. Supports multiple preset voices and custom voice design. Use when 用户提到 语音合成、文字转语音、TTS、朗读、读出来、生成语音、 生成音频、文本转音频、配音、念出来、小米语音、MiMo 语音、小米 TTS。.
byok-custom-model
Register a custom LLM endpoint with your own API key for chat in Starchild. Use when adding a personal Anthropic, OpenAI, Grok, Qwen, DeepSeek, Meta (Muse Spark), NEAR AI, or Venice key as a chat model (e.g. add my Claude key, register DeepSeek, use Muse Spark 1.1).
deepseek-vision
MUST use when the user sends or asks about images, photos, screenshots, pictures, audio, video, or mixed media documents, including requests to OCR/read text from an image. Route all media through Xiaomi MiMo V2.5 (mimo-v2.5) and mimo-v2.5-asr via scripts/mimo.py; never use local OCR, viewimage, native vision…
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…