agentleFS
finograficfinografic/ai-skills/skills/maintain-agents/SKILL.md

maintain-agents

Skill0 starsChanged 3 months agoMITView on GitHub

Review and prune the ## Learned User Preferences and ## Learned Workspace Facts sections in AGENTS.md. AI IDEs (Cursor, Copilot) append entries to these sections automatically — they accumulate stale, basic, or redundant items over time and need periodic de-slopping.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 2

    1. finografic/ai-agent-config★ 0100% same
    2. finografic/ai-agent-pipeline★ 0100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    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 repository4 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    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:

    After a long project phase — stale facts accumulate

    line 19
  2. Distinctive:

    Read the full file. Identify every bullet in:

    line 23
  3. Distinctive:

    Look for entries that overlap or repeat the same idea in slightly different words.

    line 56
  4. Distinctive:

    Ask the user to confirm or adjust before writing anything.

    line 67
  5. Distinctive:

    After confirmation, edit AGENTS.md in place:

    line 71
  6. Distinctive:

    Do not reorder the surviving entries unless the user asks.

    line 77
  7. Distinctive:

    Keep the message factual. No bullet lists in the commit body — it is a housekeeping commit.

    line 86

≈ “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

Maintain AGENTS.md — Prune Learned Sections

AI IDEs (especially Cursor) append bullet points to ## Learned User Preferences and ## Learned Workspace Facts after each session. Over time these sections bloat with entries that are basic, redundant, outdated, or already covered by instruction files. This skill describes the review-and-prune procedure.

When to run

  • User asks to "clean up", "review", "de-slop", or "prune" AGENTS.md
  • Either Learned section exceeds ~8 entries
  • After a long project phase — stale facts accumulate

Step 1 — Read the current AGENTS.md

Read the full file. Identify every bullet in:

  • ## Learned User Preferences
  • ## Learned Workspace Facts

Note: these sections may be absent (nothing to do), or there may be only one of them.

Step 2 — Evaluate each entry

Apply these criteria to every bullet. Be decisive — when in doubt, remove.

Remove if the entry…

  • States something any competent TypeScript / Node developer would know without being told (e.g. "prefer const over let", "use async/await", "always handle errors")
  • Duplicates content already covered by a linked instruction file in ## Rules — Global (read the instruction file briefly if unsure)
  • References a file, function, flag, or feature that no longer exists in the repo
  • Describes a one-off workaround for a bug that has since been fixed
  • Is a generic best practice that applies to every JS/TS project equally
  • Restates something already explicit elsewhere in AGENTS.md (e.g. in Rules sections)
  • Is a formatting micro-rule already covered by the linter / oxfmt config

Keep if the entry…

  • Documents a non-obvious project-specific pattern that would trip up an AI without knowing it
  • Records a user preference that contradicts a common default (e.g. unusual naming convention)
  • Describes a tooling quirk specific to this workspace not derivable from the code
  • Notes a subtle distinction — e.g. two similar utilities that are easy to confuse
  • Has an implied history: it was written because an AI got it wrong, and it would get it wrong again

Step 3 — Identify consolidation opportunities

Look for entries that overlap or repeat the same idea in slightly different words. Merge them into a single, more precise bullet. Fewer, sharper entries beat many vague ones.

Step 4 — Present the proposed changes

Before editing, list:

  1. Remove — each entry to delete, with a one-phrase reason
  2. Merge — any consolidations, showing before → after
  3. Keep — count of entries that survive unchanged

Ask the user to confirm or adjust before writing anything.

Step 5 — Apply the changes

After confirmation, edit AGENTS.md in place:

  • Delete removed bullets
  • Replace merged bullets with the consolidated version
  • Preserve all other content exactly (headings, order, other sections)

Do not reorder the surviving entries unless the user asks.

Step 6 — Commit

git add AGENTS.md
git commit -m "chore: prune Learned sections in AGENTS.md"

Keep the message factual. No bullet lists in the commit body — it is a housekeeping commit.

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.