agentleFS
Sign inStart free
mindmuxaimindmuxai/brain.md/skills/brain-ingest/SKILL.md

brain-ingest

Skill594 starsChanged 3 months agoApache-2.0View on GitHub

The process for digesting a conversation, document, or research result, classifying it, and writing it down as brain content (a root-page update or a new/updated page) through the `brain` CLI.

3 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 1

    1. VGBRO/claude-skills★ 198% same

    Wrote their own version · 2

    1. nikolajhh2008-svg/brainwarden★ 2same name
    2. t0ddharris/compound-marketing★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 2

    Distinctive

    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:

    Every read and write below is a brain CLI subcommand (list-pages / read-page / read-root to read.

    line 12
  2. Very specific:

    Not found → create it: read the brain-page skill, then brain create-page, then fill in the real understanding with brain update-truth.

    line 32
  3. Distinctive:

    Avoid duplication: don't scatter one topic across multiple pages.

    line 33
  4. Very specific:

    All writes go through the brain CLI — see brain-page for exact commands.

    line 37
  5. Very specific:

    Write the body in the user's working language.

    line 39
  6. Distinctive:

    Prefer less but accurate: Only capture knowledge that still matters in six months and is hard to reconstruct from the code / git.

    line 54

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

brain-ingest

This skill is about "turning scattered input into structured brain knowledge". The input can be a conclusion from a conversation, an external document, a piece of research, or the decisions behind a set of code changes. The goal is to land that knowledge in the right place, the right category, and the right structure — and to write it through the brain CLI so it is correct by construction.

Reach for this process not only when explicitly asked to "ingest" something. The brain is the project's persistent memory (see BRAIN.md), so trigger it proactively whenever knowledge crystallizes — a decision settled in discussion, a requirement or constraint agreed, an insight that will still matter in six months. Capture it as it surfaces rather than waiting to be told.

Every read and write below is a brain CLI subcommand (list-pages / read-page / read-root to read; create-page / update-truth / … to write). The command details and the category taxonomy live in the brain-page skill — read it before creating or modifying any page.

NEVER hand-edit any file under the brain directory. All reads and writes MUST go through the brain CLI. Manual edits are unsupported and illegitimate. There is no validator and nothing at the file layer can catch a bad manual edit; correctness is guaranteed only by going through the CLI, so a hand edit silently breaks the brain's invariants.

Process

1. Break down the input

Split the input into individual atomic knowledge points. A knowledge point = one judgment / fact / decision that stands on its own. Ignore purely procedural chatter.

2. Place each knowledge point (see "Choosing where to write" in BRAIN.md)

  • Changes the project's overall positioning / architecture / stack / roadmap → rewrite the corresponding root page with brain update-root <slug> (background / architecture / flow / mindmap / stack / roadmap). The stdin body must start with an authored # Title and include real content after that heading. A missing slot is created by that write.
  • About a specific entity (a decision / concept / person / organization / reference / sub-project) → create or update a page.
  • It's common for one knowledge point to touch both — write to both sides.

3. Decide between "create" and "update existing"

  • First run brain list-pages (and brain read-page <id> on likely hits) to look for an existing page on the same topic.
  • Found → update it: brain update-truth --id <id> (rewrites compiled_truth and appends a timeline entry atomically), or brain append-timeline for a new piece of evidence that doesn't change the conclusion.
  • Not found → create it: read the brain-page skill, then brain create-page, then fill in the real understanding with brain update-truth.
  • Avoid duplication: don't scatter one topic across multiple pages.

4. Write

  • All writes go through the brain CLI — see brain-page for exact commands. Never hand-edit frontmatter or timelines.
  • Root pages are rewritten wholesale via update-root; they have no timeline. Start the body with # Title. An older root page whose stored body has no H1 may omit that heading on its next write; the stored frontmatter title is inserted.
  • Write the body in the user's working language; keep technical identifiers (ids, slugs, field names, paths) verbatim. Pass --summary in that same language on create-page and update-truth (the creation entry included). The CLI will not insert an English timeline sentence for you.
  • Connect related pages explicitly with [[page-id]].

5. Verify

The write commands reindex for you, and correctness is guaranteed by construction — there is no validator to run. To optionally confirm that every [[page-id]] resolves:

node <brain-page-bundle>/bin/brain.mjs lint-links

The CLI makes "compiled_truth changed but the timeline entry was forgotten" structurally impossible; the only way to break it is a hand edit, which is why you must never make one.

Principles

  • Prefer less but accurate. Only capture knowledge that still matters in six months and is hard to reconstruct from the code / git.
  • Every compiled_truth rewrite carries its own timeline trace — update-truth guarantees this, keeping the chain of evidence unbroken.
  • Keep the category boundaries clear — the taxonomy lives in the brain-page skill.

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