Memory

Agentic skills; for hako code and other agents.

Install

npx -y skills add mithraeums/skills --skill memory

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Persistent file-based memory for any agent. Keep durable facts across sessions in a MEMORY.md index + one-fact-per-file, with rules for what to save, how to recall, and how to avoid stale/duplicate memories. Use when the user shares a durable preference, a project constraint, or says "remember this" / "from now on" — or at session start to recall. Host-agnostic; no database, just markdown. Trigger: /memory, "remember", "don't forget", "save this", "what do you know about".

SKILL.md

2.5 KB, 488 tokens by cl100k_base, as published. Nobody here has run it

memory — durable knowledge in markdown

What this is

A portable long-term memory: a memory/ folder where each fact is one small file, indexed by a MEMORY.md you load at session start. No runtime — just files any agent can read and write.

Layout

memory/
├── MEMORY.md              ← index: one line per fact, loaded every session
└── <slug>.md              ← one fact per file, with frontmatter

Each fact file:

---
name: <short-kebab-slug>
description: <one line — used to judge relevance on recall>
type: user | preference | project | reference
---
<the fact. link related facts with [[other-slug]].>

MEMORY.md line: - [Title](slug.md) — short hook. One line each. Never put fact bodies in the index.

What to save

user — who they are: role, stack, expertise, hard preferences.
preference — how they want you to work (a correction, a confirmed approach) + why.
project — goals/constraints not derivable from the code or git history; convert relative dates to absolute.
reference — pointers to external resources (URLs, dashboards, tickets).

What NOT to save

Anything the repo already records (code structure, past fixes, git history).
Things that only matter to the current conversation.
If asked to remember something obvious, ask what was non-obvious about it and save that.

Recall

At session start (or when a task touches a known area), read MEMORY.md; pull the fact files whose description looks relevant. Treat recalled facts as point-in-time — if one names a file/flag/function, verify it still exists before relying on it.

Hygiene

Before saving, check for an existing file that covers it — update, don't duplicate.
Delete facts that turn out wrong.
Link related facts with [[slug]]; a link to a not-yet-written slug is fine.

Notes

Pure markdown, no deps — works in any agent. Pairs with handoff (session-state) and corp (give agents a shared STATE.md).

Gives 0 of the 12 instructions most memory context skills give in 488 tokens

Counted across 674 of the 847 authors here whose files we hold, read 2026-08-06

inform the user when setup is completein 21 of 674, across 6 files
confirm the draft with the user before writingin 21 of 674, across 6 files
update the agent skills block in place if it existsin 21 of 674, across 6 files
present findings to the userin 20 of 674, across 5 files
write the three docs files from seed templatesin 20 of 674, across 5 files
ask the user about each decision one at a timein 19 of 674, across 4 files
edit CLAUDE.md if it existsin 18 of 674, across 3 files
explore current repo statein 18 of 674, across 3 files
do not overwrite user edits to surrounding sectionsin 18 of 674, across 3 files
back up the original file before overwritingin 16 of 674, across 8 files
keep the memory index under 200 linesin 15 of 674
Provide actionable steps and verificationin 13 of 674, across 2 files

Said here and by no other author read

pull relevant fact files on recall
delete facts that are wrong
ask what was non-obvious before saving obvious facts

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.