agentleFS
KeremozdemirraKeremozdemirra/agent-kit/skills/hafiza-guncelle/SKILL.md

hafiza-guncelle

Skill0 starsChanged 54 days agoMITView on GitHub

Oturumda öğrenilenleri kalıcı hafızaya işler: yeni tercihler, proje bilgileri, kısaltmalar, kararlar. \"Bunu hatırla\", \"hafızaya ekle\", \"bunu not al\", \"bir daha sormana gerek kalmasın\" dendiğinde veya uzun bir oturumun sonunda kullan. Also triggers on \"remember this\", \"add this to memory\", \"save this so I don't have to repeat it\", \"note that down for next time\". Tek seferlik görev detayı, geçici durum veya şifre ve API anahtarı gibi hassas veri için kullanma.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Put it in the right place: Write into the correct section.

    line 41
  2. Distinctive:

    Write what can be checked: "reports as PDF" ✅ / "the user likes tidiness" ❌

    line 50
  3. Distinctive:

    Never: write sensitive data (passwords, API keys, ID numbers, bank details).

    line 51
  4. Distinctive:

    If there is no memory file: ask where it should live, then create it.

    line 59

≈ “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

Updating memory

The point: never having to explain the same thing twice.

What goes in, what does not

Goes in (things that will be true again):

  • Durable preferences ("always deliver reports as PDF", "never use emoji")
  • A project's name, purpose, stack, and status
  • People and their roles
  • Abbreviations, code names, in-house jargon
  • Decisions taken, and why
  • Recurring workflows
  • Corrections ("do not do X again, because Y")

Does not go in: one-off task detail, anything already written in a file (link to it instead), transient state, and guesses (if you are not sure, ask; do not invent and write it down).

Steps

  1. Read the existing memory (CLAUDE.md, if there is one) so nothing is written twice.
  2. Scan the session for: the moments the user corrected you (a preference signal), sentences like "always / never again / we usually / here we do it this way", any name, abbreviation or project they explained that is not recorded, and the decisions taken.
  3. Propose, then write:
    ## Proposed additions
    § Section → "the line to add"   [why: you said this in the session]
    ## Proposed updates
    § Section → old: "..." → new: "..."
    Get approval, then apply.
  4. Put it in the right place. Write into the correct section; do not disturb the existing structure.
  5. Prune. Once the file passes 200 lines: delete lines that are no longer true, merge duplicates, and move bloated detail into a separate file with a link. Show what you are removing before removing it.

Rules

  • Keep it short. Memory is read every session; if it bloats, it becomes expensive and ineffective at the same time.
  • Write what can be checked: "reports as PDF" ✅ / "the user likes tidiness" ❌
  • Never write sensitive data (passwords, API keys, ID numbers, bank details).
  • Quote the heredoc delimiter. Memory prose carries backticked code spans and dollar signs; an unquoted <<EOF runs the backticks as commands and writes the file with the span replaced by empty output while reporting success. Write through <<'EOF' and pass any variable the script needs through the environment or a placeholder replaced afterwards.
  • Date every entry added to the decision log.

If there is no memory file: ask where it should live, then create it.

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.