本文へ移動
cccskills
無料GitHub で公開

m0

Use when the user says 'm0', 'm0 status', 'start m0', 'cross-tool memory', 'operational thread', or 'where is my memory stored', or when the M0 daemon or MCP wiring needs attention.

インストール方法を見る

含まれるファイル(3)

  • SKILL.md3.0 KB
  • scripts/m0_mcp.py15.3 KB
  • scripts/m0_server.py23.6 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

/m0 — Cross-Tool Agent Memory

M0 keeps one local operational thread per project: what was done, what was verified, what is next. Any tool can write to it and read it back, so a session that starts cold can continue work another tool began.

The live M0 is the Rust daemon coco-m0d, serving 127.0.0.1:8000 under the launchd job com.coco.m0-backend. The write path is /m0-remember, the read path is /m0-recall, and the session handoff flow is /m0-handoff.

The token is read from ~/.coco/m0-token on each call; never print it.

Quick Reference

M0="${COCO_M0_URL:-http://127.0.0.1:8000}"     # daemon base URL

# Is it up? (health needs no token)
curl -s "$M0/api/health"

# Is the launchd job loaded?
launchctl list | grep com.coco.m0-backend

# Everything else needs the token, for example the recent thread:
curl -s -H "Authorization: Bearer $(cat ~/.coco/m0-token)" \
  "$M0/api/brain/thread?project=my-project&limit=5"

There is no offline path: if the daemon is down, say so and stop; do not write anywhere else.

Sub-commands

/m0 status — is memory working

M0="${COCO_M0_URL:-http://127.0.0.1:8000}"
curl -s "$M0/api/health"
launchctl list | grep com.coco.m0-backend

Report to the user, in this order:

  1. Daemon — reachable at $M0 or not.
  2. launchd — whether the job com.coco.m0-backend is loaded and running.

If either is missing, tell the user and stop. Do not work around it.

/m0 start — you do not start it

The daemon runs under launchd; if it is down, tell the user and stop (restarts and deploys are the owner's).

/m0 mcp — the MCP tool

The daemon's MCP binary coco-platform-mcp exposes one tool, m0, with modes recall, context, wip, handoff, and remember. Call it directly — it is the lowest-friction way to make memory habitual.

/m0 verify — prove it works

curl -s "$M0/api/health"

A healthy response means the daemon is serving. Quote the output rather than summarising it.

/m0 privacy — what leaves the machine

Nothing. The daemon binds to loopback, there are no outbound calls in the request path and no telemetry. Text fields are capped at 400 characters and redacted by the daemon before storage.

When to reach for this

  • Use /m0-recall at the start of a session, or when picking up work started elsewhere, before asking the user to repeat context.
  • Use /m0-remember after a step lands, a decision is made, or something is verified.
  • Use /m0-handoff before context is compacted or a session ends.
  • Use /m0 (this skill) for the plumbing: daemon, token, wiring.

M0 answers "where were we and what is next". For deeper knowledge-graph retrieval see the comparison with the cognee bundle in systems/m0/README.md before choosing.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Train and optimize AI agents using Microsoft's Agent Lightning framework with reinforcement learning. Use when setting up agent training, instrumenting agents with tracing, configuring LightningStore, implementing reward functions, or optimizing prompts with RL/APO algorithms.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

Post-run self-evaluation system that scores agent output on correctness, clarity, actionability, and conciseness. Use after /team runs, skill executions, or when explicitly asked to evaluate output quality.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

Create AI marketing videos for ads, promos, product launches, and brand content. Models: Veo, Seedance, Wan, FLUX for visuals, Kokoro for voiceover. Types: product demos, testimonials, explainers, social ads, brand videos. Use for: Facebook ads, YouTube ads, product launches, brand awareness. Triggers: marketing video, ad video, promo video, commercial, brand video, product video, explainer video, ad creative, video ad, facebook ad video, youtube ad, instagram ad, tiktok ad, promotional video, launch video

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

Use when building AI features into a product: LLM integration, RAG pipelines, guardrails, streaming, AI UX, prompt engineering, or AI cost control. Treats prompts as code and validates every model output.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

Your AI research and engineering brain trust. 59 named personas across 8 cells covering frontier labs, applied product, model architecture, reasoning/RL/agents, alignment and interpretability, theory and science of DL, multimodal and…

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

Use when designing a new REST or GraphQL API, reviewing an API spec before implementation, setting team API standards, or migrating REST to GraphQL. Covers resources, HTTP semantics, pagination, error handling, and pitfalls.

日本語の概要は準備中です。原文の説明を表示しています。

coco-research/coco5252026年10月10日 更新

coco-research のスキルをすべて見る

このスキルの問題を報告する