agentleFS
Sign inStart free
mnemoversemnemoverse/claude-plugin/skills/memory-onboard/SKILL.md

memory-onboard

Skill0 starsChanged 37 days agoMITView on GitHub

Seeds long-term Mnemoverse memory with a project's durable conventions and decisions, taken from its README, CLAUDE.md, AGENTS.md and contributing notes, one fact per memory and without saving what memory already holds. Use when the user starts using memory on an existing project, says \"onboard this repo\", \"learn this project\", or \"seed memory from the docs\", or when memory has nothing about a project the user works on often.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Look for the tools memory_stats, memory_read, memory_list_recent and memory_write.

    line 10
  2. Very specific:

    Call memory_stats to see the existing domains.

    line 14
  3. Very specific:

    If the project has a domain, call memory_list_recent with that domain and follow the cursor to the end to see everything stored there.

    line 15
  4. Distinctive:

    Leave out anything transient (open tasks, this week's status), anything that is only a link with no content, and any secret, key, token or …

    line 29
  5. Very specific:

    Use the project's domain from step 1, or project:<repository-name> if there was none.

    line 38

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Onboard a project into memory

Check first

Look for the tools memory_stats, memory_read, memory_list_recent and memory_write. If they are missing, memory is not connected: say so in one sentence and use the bundled setup skill if it is installed, otherwise https://mnemoverse.com/docs/api/install-by-host.

1. See what is already there

  • Call memory_stats to see the existing domains. If the project already has one, such as project:<repository-name>, reuse that exact spelling.
  • If the project has a domain, call memory_list_recent with that domain and follow the cursor to the end to see everything stored there; otherwise call memory_read with the project name. Facts already in memory are not written again.

2. Read the project's own documents

In this order, where they exist: README.md, CLAUDE.md, AGENTS.md, CONTRIBUTING.md, and any architecture or decision records the README points to.

Take only what stays true for months:

  • how to build, run and test the project, with the exact commands;
  • code style and naming conventions that are written down;
  • decisions with their reasons ("we use X because Y");
  • where things live: the main folders and what each is for;
  • who owns what, if the documents say so.

Leave out anything transient (open tasks, this week's status), anything that is only a link with no content, and any secret, key, token or connection string, even an example one.

3. Show the list before writing

Present the facts as a short numbered list, each with the file it came from. Write only after the user agrees, and drop the ones the user strikes out.

4. Write

  • One memory_write per fact, content as a self-contained statement that names the project and the source file.
  • Use the project's domain from step 1, or project:<repository-name> if there was none.
  • A write can be refused when memory already holds something too close to it in the same domain. If step 1 found the same fact, count it as already known; if the fact is genuinely different, write it once more with only what is new in it, and report any fact that is still refused.

5. Report

Say how many facts were saved, how many were already known, and which files were read. Suggest the memory-search skill for using them.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free