agentleFS
finolabsfinolabs/vibecode/skills/update/SKILL.md

update

Skill4 starsChanged 14 days agoMITView on GitHub

Update finolabs-vibecode with the user's permission and walk them through what the new version asks of them, show what is new without changing anything, target the latest or a chosen released version, pin or unpin a version, skip a release, or stop update notices. Use when the user requests a plugin update, asks whether a newer release is available or what a new version changed, wants a specific or older version, chooses skip this version, or says never ask again.

63 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use end, create and 8 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.
  • torarnvSkill

    end

    Wrap up a work session. Updates memory files, logs decisions and lessons, appends to the session log.

    Also in 3 projects that use this

  • MicrosoftSkill

    create

    Also in 4 projects that use this

  • bitwize-music-studioSkill

    tutorial

    Provides interactive guided album creation for new users. Use when the user is new to the plugin or asks for a walkthrough of the album creation process.

    Also in 3 projects that use this

  • MicrosoftSkill

    note

    Use this skill to record progress, partial findings, hypotheses, and page-anchored evidence while investigating a long document. Call Note aggressively — any time you read a page and confirm, narrow, or update your understanding of the user's question. Triggers include "jot that down", "save this finding", "note that page 7 says X", or (more often) implicit internal signals — you just learned something and want it recoverable later. Notes persist in the session and Review can select among them. Never paraphrase notes back from memory; write them as you go.

    Also in 4 projects that use this

  • kv0906Skill

    inbox

    Quick capture mode (with content) or batch processing mode (no content). GTD-style routing of inbox items to proper folders. Use "/inbox thought to capture" or "/inbox" to process.

    Also in 4 projects that use this

  • jholhewresSkill

    backup

    Backup and restore — files, databases, and system backups

    Also in 3 projects that use this

  • +1 more from microsoft

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 61 edited versions we read keep these.

  • 0

    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:

    When the user has gone through that installation's list, run onboard seen <installed version> with the same re-resolved helper and run …

    line 31
  2. Very specific:

    When the user explicitly chooses to hold a version, run python3 -I <update_check.py> pin <version>.

    line 37
  3. Very specific:

    Never put response text into a shell command: versions must match v?X.Y.Z with decimal digits and dots only, and item ids must be ones …

    line 41

“Common” is measured: kept by at least 6 in 10 of the 61 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

Update the plugin

Run only at the user's request. A session-start notice tells the agent to let the user know a newer version is available, or that the installed version has upgrade notes, and that they can type /update (on Codex $update). Only the user can invoke this skill; a notice is not permission to update. An earlier yes to checking or opening this skill is not a yes to installing an update or to any setup step.

The helper and its shared core are in this skill's own scripts/ folder. Resolve <update_check.py> there. It uses fnl_metrics.state_dir() for all state, independently of metrics consent. It reads the plugin's own manifest when installed with the plugin; a standalone skill has no manifest, so its installed version is unknown. Use the host's installed-plugin listing to establish that version rather than guessing it. Command output and states are in references/onboarding.md; host commands and privacy are in references/hosts.md.

Arguments: $ARGUMENTS

Show the facts, then ask

Questions below ask through the host's native question tool when the session has one. Ask in text when the session has no such tool, the tool is refused or fails, nobody can answer (a delegated or headless run records the open decision and never assumes or invents an answer), or the decision is free text. The enforcer is agent judgment, so this is advisory.

Advisory (agent judgment): Talk to the user in the language they are using, including questions and option labels; keep command names, flags, file paths, skill names and other technical names unchanged. Script output stays English; report it to the user in their language (advisory, agent judgment).

  1. Run python3 -I <update_check.py> status. This reads local state only. Show installed, available (the cached latest stable release), whether it is newer, whether the cache is stale, any pinned version, and any skipped versions or never-ask choice. A null value is unknown, not evidence that the plugin is current. Empty output means the check is off or a read failed: report that limitation. Do not fetch a release manually or claim a stale cache is current. The startup hook schedules a background refresh after 24 hours; the result is read on a later startup. For an update, first find the target: target alone for the latest stable release (the default), or for /update <version> the evidence form in references/onboarding.md. A refusal ends the update; report its reason. For a rollback, show the lost: lines first.
  2. If the user only asked for status, finish with those facts. For an update, report the host's native command for the installed identifier, its status from references/hosts.md and any restart or trust step. Ask for a clear yes to running that command now, through the host's native question tool when the session has one. Offer neutral Yes and No choices with none preselected or marked recommended; a default, a timeout or an auto-selected option never counts as the yes. Wait for the answer. A skip, silence or ambiguous answer grants no permission.
  3. After that yes, run only a verified native update command for the matching host and installed identifier. Report the actual result. An error is an error, not a successful update. Verify the resulting installed version with the host's listing; a still-running session's manifest can describe the old loaded copy.
  4. Go straight on to the checklist (step 5); the user is never sent to a second command. First re-resolve <update_check.py> inside the newly installed copy that the host listing reports, and run its status: its installed must equal the verified version. If it does not (the host keeps the old copy loaded until a restart), stop here, say a restart is needed and that /update whats-new or /update continues afterwards, and do not run onboard seen.

The checklist

  1. Run python3 -I <update_check.py> whats-new and show it. For each [open] or [unknown] item in the order given, offer its setup step: invoke setup with that /setup add <feature>, which asks its own consent, preview and neutral yes exactly as it always does. Nothing is set up without the user's own answer there. When the user chooses to skip an item, run onboard skip <id>. Items without a setup step are told, not offered.
  2. When the user has gone through that installation's list, run onboard seen <installed version> with the same re-resolved helper and run whats-new again to verify. This clears the session-start line.

/update whats-new runs step 5's command only: it shows the checklist and changes nothing, installs nothing, offers no setup step and never calls onboard.

Pin, skip a version or stop asking

When the user explicitly chooses to hold a version, run python3 -I <update_check.py> pin <version>; unpin removes it. While a pin is set the newer-release notice is silenced and status shows pinned; the pin does not change the installed plugin.

When the user explicitly chooses skip this version, use the validated available version from status and run skip <version>. When the user explicitly chooses never ask, run never. These commands are local and silent; read status again to verify the choice before reporting success. Skip affects that release only; never-ask suppresses the release notices, but the background check remains on.

Never put response text into a shell command: versions must match v?X.Y.Z with decimal digits and dots only, and item ids must be ones whats-new printed.

Privacy and off switch

The automatic check makes one fixed HTTPS GET to the GitHub /releases/latest endpoint at most once a day and sends nothing about the user; nothing in this skill adds a network call. FNL_UPDATE_CHECK=off stops the check, the notices and these helper commands. Details are in references/hosts.md and PRIVACY.md.

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.