FAQ: Common questions about skillshare

[runkids]

A collection of frequently asked questions. If your question isn't here, feel free to open a new Q&A discussion!

---

How is skillshare different from npx skills?

The core difference is architectural:

• npx skills uses an imperative model — each add is a standalone operation. You select agents, choose copy/symlink, confirm, every time. There's no persistent config.

• skillshare uses a declarative model — you define your desired state in config.yaml (targets, sync mode), then skillshare sync reconciles reality to match. One command, no prompts, deterministic.

Practical consequences:

Scenario	npx skills	skillshare
Install a skill	Select agents interactively	install + sync (uses config)
Remove from one agent	May delete source, breaking others	Only removes that target's symlink
New machine	Re-run every install	git clone config + sync
Pull edits back from a tool	Not possible	skillshare collect <target>
Web dashboard	None	skillshare ui
Runtime dependency	Node.js + npm	None (single Go binary)

Full comparison: skillshare.runkids.cc/docs/guides/comparison

---

Which AI CLIs does skillshare support?

40+ targets including:

Officially tested: Claude Code, Cursor, Codex CLI, OpenCode, OpenClaw, Gemini CLI, Amazon Q, Windsurf, and more.

Custom targets: Any tool that reads skills from a directory:

skillshare target add my-tool ~/.my-tool/skills

Full list: skillshare.runkids.cc/docs/targets/supported-targets

---

What's the difference between merge and symlink sync modes?

• merge (default): Creates individual symlinks for each skill inside the target directory. Your target can still have its own local skills alongside synced ones.

• symlink: Symlinks the entire target directory to the source. Simpler, but the target can only contain synced skills.

Most users should stick with merge. Use symlink only if you want the target to be an exact mirror of the source.

---

Does skillshare send any data / telemetry?

No. Zero telemetry, zero analytics, zero network calls (except when you explicitly use install, search, push, or pull).

Your skill setup stays entirely local. There is no central registry.

---

Can I use skillshare for project-scoped (per-repo) skills?

Yes. Project mode keeps skills inside .skillshare/ in your repo:

cd your-project
skillshare init -p
skillshare install org/skills -p
git add .skillshare && git commit -m "Add project skills"

Teammates clone the repo, run skillshare sync, and every AI tool picks up the project skills.

Docs: skillshare.runkids.cc/docs/concepts/project-skills

---

How do I share skills across my team / organization?

Two approaches:

A. Tracked repo (recommended for teams):

skillshare install github.com/your-org/team-skills --track
skillshare sync

Everyone installs the same repo. Updates via skillshare update --all.

B. Project mode (for per-repo skills):

skillshare init -p
# add skills to .skillshare/skills/
git push

Teammates get skills automatically when they pull the repo.

Guide: skillshare.runkids.cc/docs/guides/organization-sharing

---

I have skills in multiple AI CLIs already. Will I lose them?

No. The migration process is non-destructive:

skillshare init              # creates config
skillshare backup            # backs up all targets first
skillshare collect --all     # copies skills to source
skillshare sync              # replaces copies with symlinks

backup creates a snapshot before any changes. You can always skillshare restore <target> to go back.

---

Does it work on Windows?

Yes. Windows uses NTFS junctions instead of symlinks — no admin privileges required.

irm https://raw.githubusercontent.com/runkids/skillshare/main/install.ps1 | iex

Troubleshooting: skillshare.runkids.cc/docs/troubleshooting/windows

---

How do I update skillshare itself?

skillshare upgrade

Or with Homebrew:

brew upgrade skillshare

---

Don't see your question? Open a new discussion — happy to help!