agentleFS
relote-comrelote-com/omarchy-context-engine/skills/context-engine/SKILL.md

context-engine

Skill4 starsChanged 11 days agoMITView on GitHub

Shared long-term memory for this Omarchy computer, held in Context Engine and reached through the `context-engine` MCP server (tools brief, find, get, write, edit) or the `ce` command. Use at the start of every session to load what changed, before answering anything about past decisions, people, projects or why this machine is set up the way it is, and after making a decision or changing the system so the next agent — whichever one it is — knows. Triggers: remember, recall, "what did we decide", "why is X like this", "last time", notes, decisions, tasks, my setup, my config history.

9 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use andreessen, business-investment-advisor and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Hussain-ElfalahSkill

    andreessen

    Also in 4 projects that use this

  • Hussain-ElfalahSkill

    business-investment-advisor

    Also in 4 projects that use this

  • csiddhant796-blipSkill

    karpathy-coder

    Use when writing, reviewing, or committing code to enforce Karpathy's 4 coding principles — surface assumptions before coding, keep it simple, make surgical changes, define verifiable goals. Triggers on "review my diff", "check complexity", "am I overcomplicating this", "karpathy check", "before I commit", or any code quality concern where the LLM might be overcoding.

    Also in 4 projects that use this

  • +9 more from Hussain-Elfalah

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Prefer the MCP tools on the context-engine server:

    line 15
  2. Very specific:

    If the MCP tools are not available in this session, the ce command does the same over HTTP:

    line 25
  3. Very specific:

    Session start: Call brief once, before your first reply.

    line 39
  4. Distinctive:

    Save when something is decided or learned,: not at the end of every turn:

    line 45

“Common” is measured: kept by at least 6 in 10 of the 8 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

Context Engine — the memory every agent on this machine shares

This computer runs Omarchy. The person can launch any of several coding agents (omarchy agent), and all of them are connected to the same Context Engine. What you save, the next agent reads. What the last agent saved, you can read now. Treat the engine as the machine's memory, not your own notes.

How to reach it

Prefer the MCP tools on the context-engine server:

ToolUse it to
briefStart a session: which engines you reach, what each is for, whether you may save there, what changed since the last session, and what never to store.
findSearch (q) or list recent memories, filtered by type, tags, since. Bodies come back inline.
getOpen one memory by ref, e.g. decision::hyprland-gaps.
writeSave a new memory in one engine (required): note, decision (needs fields.rationale), task, person, company, meeting. Always tag it with this machine (below).
editChange a memory that already exists instead of writing a second one.

If the MCP tools are not available in this session, the ce command does the same over HTTP:

ce brief                     # what changed
ce find "hyprland gaps"      # search
ce get decision::hyprland-gaps
ce remember "Title" --decision --why "because…" --tag hyprland   # tags the machine itself
echo "longer note…" | ce remember --title "Title"
ce engine                    # which engine new memories go to

What to do

  1. Session start. Call brief once, before your first reply. Mention anything in it that bears on the request; don't recite it.
  2. Before answering from assumption, find first when the question touches the person's past: "why is…", "what did we…", "last time", a named person or project, or a config choice you did not make. Cite the memory you used by its ref.
  3. Save when something is decided or learned, not at the end of every turn:
    • A choice with a reason → decision with fields.rationale.
    • A fact worth knowing later → note.
    • Work someone still has to do → task. find first; if a memory about the same thing exists, edit it.
  4. Changing this Omarchy system (anything under ~/.config/hypr, ~/.config/omarchy, terminal configs, themes, keybindings, installed packages): after the change works, save a decision titled after what changed, with the old value, the new value, the file, and the person's reason in rationale. Tag it omarchy plus the area (hyprland, bar, theme, keybinding, package). This is how the next agent answers "why is my gap 5?" months later, and how the person's other machines learn the same preference.
  5. Say which machine. The person may run Omarchy on several computers that share one engine. Every write carries the tag machine:<hostname>, with the hostname exactly as uname -n prints it on this computer (run it once per session), next to your other tags: tags: ["omarchy", "hyprland", "machine:air"]. ce remember adds it for you. When a memory you found carries another machine's tag, say so if it matters ("on your desk machine, gaps are 10").
  6. Use the default engine (ce engine prints it) unless the brief or the person says another one fits better: a client's engine for that client's work, a team engine for team decisions.

What never to store

Secrets — API keys, tokens, passwords, private keys — never; store where one lives, not what it is. Anyone's private data (health, finances, home address, contact details) only when the person asks you to keep that exact thing. The brief repeats this rule; follow the engine's wording if it differs.

Style of a good memory

  • Title says the thing: "Hyprland gaps set to 5 on the Air", not "Config update".
  • Body is a few lines of Markdown a stranger could act on. Link related memories with [[type::slug]].
  • One fact per memory. A long session yields a few memories, not a transcript.

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.