Recipe: Team Onboarding
Give new teammates the team's shared skills and project context.
Scenario
A new developer joins your team. They need:
- Organization-wide skills (coding standards, review guidelines)
- Project-specific skills (domain knowledge, architecture rules)
- Everything working across their AI tools (Claude Code, Pi, etc.)
One teammate uses Claude Code, another uses Codex, and the new hire uses Pi. All three need a review checklist that says which legacy API to avoid. Copying that checklist into each tool creates separate versions that drift as the project changes.
Keep the project's local skills, .skillshare/config.yaml and .skillshare/skills.lock.json in the project repository. The config declares remote skills and targets; the lockfile records remote skill commits. Each teammate applies those files locally. Shared instructions give the team common context, while each tool retains its own permissions and behavior.
Solution
Step 1: Create an onboarding script
Save as scripts/setup-skills.sh in your team wiki or repo:
#!/bin/bash
set -e
echo "Installing skillshare..."
curl -fsSL https://raw.githubusercontent.com/runkids/skillshare/main/install.sh | sh
export PATH="$HOME/.local/bin:$PATH"
echo "Initializing..."
skillshare init -g
echo "Installing organization skills..."
skillshare install github.com/your-org/org-skills --track -g
echo "Running security audit..."
skillshare audit -g --threshold high
echo "Syncing to all AI tools..."
skillshare sync -g
echo "Done! Run 'skillshare list -g' to see installed skills."
Step 2: New hire runs the script
curl -fsSL https://your-org.github.io/setup-skills.sh | sh
Or if the script is in the team repo:
git clone your-org/team-tools
./team-tools/scripts/setup-skills.sh
Step 3: Project-specific setup
The maintainer first follows Project Setup and commits the project config, local skills and lockfile. When the new hire clones that project:
cd your-project
skillshare install -p
skillshare audit -p --threshold high
skillshare sync -p
install -p installs the remote skills declared in .skillshare/config.yaml, using locked commits when present. audit -p reviews the project skills, including committed local skills. sync -p distributes them to the configured targets. Run these commands in order and stop if a step fails; a blocked audit needs review before syncing.
After pulling project updates, repeat this sequence. Git transfers the configuration and lockfile; it does not install missing remote skills or refresh target copies by itself. Review intentional skill updates in a PR and commit the resulting lockfile changes.
Step 4: Verify everything works
# Check global skills
skillshare list -g
# Check project skills
skillshare list -p
# Check sync status
skillshare status -p
Verification
skillshare list -gshows organization skillsskillshare list -pshows the project's local and installed remote skillsskillshare status -pshows the configured project targets are synced- Opening a configured AI tool shows the expected skills; try the review checklist against a known legacy-API example
Variations
- Dev container onboarding: If your team uses dev containers, add skillshare to
.devcontainer/DockerfileandpostCreateCommand— skills are ready when the container starts - Homebrew-based install: Replace
curl | shwithbrew install skillsharefor macOS/Linux teams - Hub discovery: Point new hires to your hub:
skillshare search --hub https://your-org.github.io/skillshare-hub.json