跳到正文
cccskills
免费在 GitHub 公开

m0-remember

Use when the user says 'm0 remember', 'record what we did', 'note for next session' or 'write a checkpoint', or after finishing a step worth handing to the next session. Appends one entry to the M0 thread (re-sending writes a second entry unless you pass the same `id`).

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md4.5 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

/m0-remember — Write to the Operational Thread

The write path for M0. One call appends one entry to the shared thread for a project.

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}"

curl -s -X POST "$M0/api/brain/checkpoint" \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $(cat ~/.coco/m0-token)" \
  -d '{"project":"acme-web","kind":"step_done",
       "text":"Fixed the token refresh race in the auth middleware.",
       "next_step":"Add a regression test for two concurrent refreshes.",
       "last_verified":"pytest tests/auth -q: 31 passed",
       "source_tool":"claude_code"}'

If the MCP tool is wired (/m0 mcp), call m0 in remember mode directly — same endpoint, fewer moving parts.

Fields

FieldRequiredWhat to put in it
projectyesThe project key. Use one stable value per project — the repository or directory name is the usual choice. Getting this wrong splits the thread.
textyesWhat happened, in one or two plain sentences. Capped at 400 characters and redacted by the daemon. Written for a reader with no other context.
kindnoDefaults to step_done. See the table below.
next_stepnoThe single next action. Concrete enough to act on without re-deriving it.
last_verifiednoWhat was actually checked, and how. A command and its result, not an impression.
session_idnoSession identifier, when known.
rolenoThe role the entry was written under, when it matters.
source_toolnoWhich tool is writing: claude_code, cococode, ambient, manual, hook, paperclip. Fill it in — it is what makes the thread legible across tools.
branch, head_shanoVersion-control position. Worth including whenever the entry is about code.

Kinds:

kindUse it for
step_doneA completed step, a fact, or a decision. The default.
compact_checkpointA session handoff — see /m0-handoff.
session_endA session closing, usually from a hook.
lane_dispatchedWork handed to a subagent or parallel lane.
lane_resultThe outcome of that work.
ambient_signalContext observed rather than reported.

An unknown kind is rejected, deliberately: a typo would create a category no reader looks in.

Procedure

  1. Resolve the project key. Use $M0_PROJECT if set, otherwise the repository or directory name. Reuse whatever earlier entries used — check with /m0-recall if unsure. Do not invent a variant.

  2. Write one entry per meaningful thing. A step that landed, a decision with its reason, a verification result. Not a running commentary.

  3. Include version-control context when the entry is about code:

    BRANCH="$(git rev-parse --abbrev-ref HEAD 2>/dev/null)"
    SHA="$(git rev-parse --short HEAD 2>/dev/null)"
    
  4. Be honest in last_verified. Put the command and its actual result there. If nothing was verified, leave it empty. A false verification claim in a memory store outlives the session that made it and misleads every later reader.

  5. A durable Lab decision or ruling is not an m0-remember. It becomes a decision record via decide new (HQ commits records); the thread gets at most a step_done pointing at it.

Good and bad entries

text:          "Fixed the token refresh race in the auth middleware: the retry
                path double-incremented the nonce."
next_step:     "Add a regression test for two concurrent refreshes."
last_verified: "pytest tests/auth -q: 31 passed"
text:          "Made some progress on auth."          # nothing to act on
next_step:     "Continue."                            # not a next step
last_verified: "Tests should pass now."               # a claim, not a check

Notes

  • Local-only. Loopback daemon, no outbound calls, no telemetry. Text fields are capped at 400 characters and redacted by the daemon.
  • There is no offline path: if the daemon is down, say so and stop; do not write anywhere else.
  • Reading back: /m0-recall. Handoffs: /m0-handoff. Plumbing: /m0.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

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.

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19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.

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19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

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19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.

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19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…

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19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.

中文简介正在准备中。当前显示原文说明。

coco-research/coco7602026年10月19日 更新

查看 coco-research 的全部技能

报告此技能的问题