skills-manager
Skill3 starsChanged 37 days agoMITView on GitHub
Keep the AGENTS.md skills tables generated from skill descriptions, and understand how a fork syncs its skills from upstream. Use after editing a skill's frontmatter, when adding or removing a skill, or when the AGENTS.md skills-table check fails.
12 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- AlomoesDataScience/claude-skillstheir own version, first seen Nov 8, 2026
- akunzai/skills-managertheir own version, first seen Oct 17, 2026
- yaojie-shen/Skills-Manager-TUItheir own version, first seen Oct 17, 2026
- joker1point/agent-skill-frameworktheir own version, first seen Oct 16, 2026
- anyproto/anytype-swifttheir own version, first seen Oct 9, 2026
- nrjdalal/zerostarteredited, first seen Oct 8, 2026
Used with
Projects with it also use tinyagi-admin, send-user-message and 5 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.
jlia0Skilltinyagi-admin
Manage and operate the TinyAGI system itself — agents, teams, settings, queue, tasks, daemon lifecycle, and source code. Use when the agent needs to: list/add/remove/update agents or teams, check queue status, view logs, start/stop/restart TinyAGI, change settings (provider, model, channels), send messages to the queue, manage tasks, retry dead-letter messages, view recent responses, modify TinyAGI source code or configuration, or perform any administrative operation on the TinyAGI platform. Triggers: 'manage tinyagi', 'add an agent', 'remove a team', 'check queue', 'view logs', 'restart tinyagi', 'change provider', 'update settings', 'create a task', 'modify tinyagi code'.
Also in 3 projects that use this
Maliot100XSkillsend-user-message
Send a proactive message to a paired user via their channel (Discord, Telegram, or WhatsApp). Use when the agent needs to notify, alert, or send an unsolicited message to a user — especially during heartbeat invocations, scheduled tasks, or when the agent wants to reach out without a prior user message in the current conversation. Triggers: 'send message to user', 'notify user', 'alert user', 'message the user on discord/telegram/whatsapp', or any need to proactively communicate with a paired sender.
Also in 3 projects that use this
bivexSkillschedule
Управление расписанием занятий с специалистами по самосовершенствованию. Используй когда пользователь спрашивает о расписании, хочет добавить/удалить/перенести занятие, посмотреть план на день или неделю.
Also in 3 projects that use this
OpenAISkillimagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas.
Also in 3 projects that use this
genomewalkerSkilltasks
Browse and resume tasks, threads, and background jobs across sessions
Also in 3 projects that use this
GoogleSkillmemory
Records and appends every user-agent customer support interaction to the memory.md log in the workspace.
Also in 3 projects that use this
Other versions12 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- nrjdalal/zerostarter★ 6596% same
Wrote their own version · 11
- jlia0/tinyclaw★ 3.6ksame name
- TinyAGI/tinyagi★ 3.6ksame name
- TinyAGI/tinyclaw★ 3.6ksame name
- anyproto/anytype-swift★ 525same name
- shenysun/skills-manager★ 243same name
- EfanWang/skills-manager★ 30same name
- yaojie-shen/Skills-Manager-TUI★ 3same name
- akunzai/skills-manager★ 1same name
- AlomoesDataScience/claude-skills★ 1same name
- ddmsolutions/claude-skills★ 0same name
1 more not listed
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 repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- add-packageSkill
- agent-browserSkill
- api-endpointSkill
- auditSkill
- codebase-mapSkill
- db-migrationSkill
- designSkill
- devSkill
- docker-testSkill
- doc-syncSkill
- fontsSkill
- gh-commitSkill
- iceboxSkill
- ignore-syncSkill
- portlessSkill
- releaseSkill
- runtime-apisSkill
- shadcn-syncSkill
- shipSkill
- ui-verifySkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 10 edited versions we read keep these.
1
Distinctive
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
- Distinctive: line 47
Check state from inside a fork with:
“Common” is measured: kept by at least 6 in 10 of the 10 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.
Files
[!CAUTION] Synced from https://github.com/nrjdalal/zerostarter. Customize this skill or remove this note to stop syncing.
Skills Manager
The AGENTS.md skills tables duplicate each skill's own description, so they are generated, never hand-kept. .github/scripts/skills-manager.ts is the maintainer, and a pre-commit hook runs it automatically whenever a skill (or the manager) changes, so the catalog never drifts.
Each skill carries its provenance in frontmatter:
source: localis authored here (this repo is the origin).source: <tool>is vendored: re-synced by re-running the tool, never hand-edited.source: https://github.com/<owner>/<repo>marks a skill a fork syncs from upstream (the CLI stamps this, with a[!CAUTION]sync note, when it scaffolds a fork).
Provenance is never rewritten by a sync: a local or vendored skill in a fork stays that way, because a wrong source is a silent, self-propagating loss of where a skill actually came from.
Regenerate the tables
The hook handles this on commit, but to run it by hand:
bun .github/scripts/skills-manager.ts # rewrite the tables from the skills
bun .github/scripts/skills-manager.ts --check # fail on drift instead of writing (the gate)Each cell is the description's summary, the sentence before Use ..., so keep every description in the <summary>. Use when <triggers> shape and edit the description, not the table. Done when --check passes.
How a fork syncs
A fork inherits its skills from the scaffold. On init and sync the CLI rebrands each skill's prose to the fork's project name (read from package.json). A skill this repo authored (source: local) becomes source: <upstream repo> with a [!CAUTION] note at the top; a skill this repo vendored keeps its tool as source and takes no note, so the fork re-vendors it the same way and it stays in that fork's Vendored table.
Provenance decides what a later sync may touch: it rewrites only the skills sourced from here, and leaves every other one exactly as the fork has it.
That note is the contract: bunx zerostarter updates a skill only while the note is intact and the body still matches what the CLI last wrote. Customize the skill or drop the note and the fork owns it, and sync names it in the summary instead of replacing it.
Prose cannot be compared directly (sync rebrands the body to the fork's name, so a synced skill never matches upstream byte for byte), so the CLI records what it wrote in .agents/skills/.sync.json: per skill, the ref it came from plus a hash of the upstream file and of the file as written. That ledger is what separates "the fork edited this" from "upstream moved". The ref is recorded because the CLI syncs a fork from main while this repo's default branch is canary, so comparing against the wrong one would report every skill canary is ahead on as drifted. A fork commits the ledger like any other synced file; this repo has none, since no skill here is synced from an upstream repo: every skill is source: local or vendored under its tool's own name.
AGENTS.md ships the same way and is tracked in the same ledger, under its own path. init rebrands it; sync hands the fork the update only while its copy is still what the CLI last wrote, and keeps an edited one. The generated skills tables are left out of that comparison, since the fork's own skills-manager rewrites them. It carries no note: editing it is the opt-out. A guide with no sync record is replaced only when it is the untouched stub an older CLI scaffolded; anything else is the fork's own work and stays. AGENTS.md is also a PRESERVE_ON_SYNC path, which is what keeps an older CLI from overwriting it.
Check state from inside a fork with:
bun .github/scripts/skills-manager.ts --outdatedwhich reports each skill as local, no upstream, vendored (<tool>), untracked (customized there, or synced before the ledger existed), or, for a tracked skill, up to date / DIFFERS from upstream, plus edited here when the fork has since changed it.
Notes
- Adding or removing a skill changes the tables; the hook regenerates them, and
--checkis the gate if it is ever bypassed. name,description, andsourceare the contract the manager reads; a missing one makes it throw rather than emit a half-built table.
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
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.

