How to Sync and Back Up Claude Skills Across Multiple Computers

Why Claude Skill Sync Breaks in Real Work

Most Claude Skills start as local experiments. You write a useful review checklist on one laptop, save a customer-reply skill in one repo, or get a teammate to send you a folder over chat. That works until you change machines, open a different project, or try to use the same workflow in another agent.

The failure mode is rarely dramatic. It is usually quieter: the old laptop has the newest version, the repo has a stale copy, a support file is missing, or two folders share almost the same name. By the time you notice, you are rebuilding the workflow from memory.

A synced skill is not a synced Markdown file. It is a complete, versioned workflow package that can be found, executed, and tested on every machine that needs it.

The Sprawl Moment: The Cost of Unbacked and Desynced Skills

A lead platform architect spends four months engineering a suite of 18 high-precision agent skills across Claude Code, Google Antigravity, and Cursor. These skills automate internal API contract validation, multi-stage AWS infrastructure rollouts, and SOC2 compliance audits. Following an unexpected macOS kernel panic and clean OS reinstallation, the architect discovers that while git repos were preserved on GitHub, all custom skills residing in ~/.claude/skills and ~/.gemini/antigravity/skills were completely wiped. Re-authoring the nuanced prompt heuristics, regular expressions, and edge-case exceptions from memory takes over 30 hours of lost development time, introducing regressions across the entire team's pull request reviews.

Prompttly is a skill manager for AI agents — one library for your skills and prompts that syncs into Claude Code, Codex, ChatGPT, and Claude and is one hotkey away on your Mac, so your setup follows you across every machine, repo, and tool.

Choose the Right Sync Method

Search intent for sync Claude Skills is informational with a practical setup need: developers and knowledge workers want to know what will work reliably without corrupting their skill library or losing multi-agent configs. The right answer depends on whether you are syncing for yourself, one repo, or an entire team.

Comparing dedicated tools to synchronize rules and prompt packages across multi-agent setups? Explore our evaluation of the best AI skill managers.

Manual Baseline: Copy the Whole Skill Folder

If you only need to move one skill today, copy the whole folder. Do not copy only SKILL.md unless the skill truly has no examples, templates, references, scripts, or assets.

# On the old machine
tar -czf claude-skills-backup.tgz -C ~/.claude skills

# On the new machine
mkdir -p ~/.claude
tar -xzf claude-skills-backup.tgz -C ~/.claude

This is a backup and migration pattern, not a long-term sync system. It is useful when setting up a new laptop, but it does not solve ongoing edits from two machines.

If you are not sure where the folders belong, read our guide on Claude skills across projects and directory scopes before moving files. For a one-time hardware change, follow the Claude Code settings transfer checklist before you wipe the old computer.

If your main risk is losing a working library before you sync it, start with our disaster recovery and backup verification workflow below so you have a tested restore path first. If your main risk is a library full of duplicates, use the Claude Skills organization guide before syncing messy folders to every machine.

Disaster Recovery: Backing Up Multi-Agent Skills Before Syncing

Backing up Claude Skills and AI agent workflows protects months of refined engineering judgment, structured quality assertions, and automated tool integrations. When developers look to backup Claude Skills, the underlying risk is not simply recovering a Markdown snippet; it is preserving the dependable, repeatable behaviors that power autonomous development workflows in Claude Code, Google Antigravity, and OpenAI Codex.

As detailed in the Anthropic Claude Code Overview, agent skills represent modular software capability packages. A complete backup must preserve 3 indispensable layers: the procedural SKILL.md manifest, all supporting execution scripts and references in subdirectories, and the granular Git commit history that explains why constraints were introduced.

Before creating a backup archive, catalog every local directory where AI agent skills and custom instructions reside across your workstation:

1. Claude Code Skills: Global skills in ~/.claude/skills/ and project-scoped skills in <repo>/.claude/skills/.
2. Google Antigravity Skills: Global skills in ~/.gemini/antigravity/skills/ and workspace skills in <repo>/.gemini/skills/. (See our Google Antigravity skills guide).
3. OpenAI Codex Skills: Global packages in ~/.codex/skills/ and repository rules in AGENTS.md. (See our Codex skills guide).
4. Cursor IDE Rules: Workspace-scoped rule manifests in <repo>/.cursor/rules/*.mdc.

How Do Different Skill Backup Strategies Compare?

Choosing the right backup strategy depends on whether you require quick point-in-time archives, collaborative team version control, or automated multi-agent synchronization across multiple operating systems.

Automating Multi-Agent Skill Backups with Shell Scripts

To create an automated, reproducible backup that captures skills across Claude Code, Google Antigravity, and OpenAI Codex, use this POSIX-compliant backup automation script. It generates a compressed, timestamped tarball and stages changes cleanly.

#!/usr/bin/env bash
set -euo pipefail

BACKUP_DIR="${HOME}/.ai-skill-backups"
TIMESTAMP="$(date +%Y-%m-%d_%H%M%S)"
ARCHIVE_PATH="${BACKUP_DIR}/ai-skills-backup-${TIMESTAMP}.tar.gz"

mkdir -p "${BACKUP_DIR}/staging"

echo "[*] Archiving Claude Code skills..."
[ -d "${HOME}/.claude/skills" ] && cp -R "${HOME}/.claude/skills" "${BACKUP_DIR}/staging/claude-skills"

echo "[*] Archiving Google Antigravity skills..."
[ -d "${HOME}/.gemini/antigravity/skills" ] && cp -R "${HOME}/.gemini/antigravity/skills" "${BACKUP_DIR}/staging/antigravity-skills"

echo "[*] Archiving OpenAI Codex skills..."
[ -d "${HOME}/.codex/skills" ] && cp -R "${HOME}/.codex/skills" "${BACKUP_DIR}/staging/codex-skills"

echo "[*] Compressing backup archive..."
tar -czf "${ARCHIVE_PATH}" -C "${BACKUP_DIR}/staging" .
rm -rf "${BACKUP_DIR}/staging"

echo "[+] Successfully created skill backup: ${ARCHIVE_PATH}"

Refer to the official Git Documentation for best practices on structuring semantic commit messages when pairing local archive scripts with remote repository tracking.

How Do You Test and Verify a Skill Backup Before Disaster Strikes?

An untested backup is merely an assumption of safety. To guarantee that your AI agent skills will perform flawlessly following a machine migration or disk failure, execute this 5-point disaster recovery checklist in an isolated sandbox environment:

1. Step 1: Extract into a Clean Staging Directory. Extract the archive into a temporary folder (e.g. /tmp/skill-restore-test) to verify archive integrity.
2. Step 2: Verify Multi-File Hierarchy. Confirm that all scripts/, references/, and examples/ directories are intact and have appropriate read/execute permissions (chmod +x).
3. Step 3: Validate YAML Manifests. Check that all SKILL.md files parse valid YAML frontmatter without syntax errors or broken indentation.
4. Step 4: Execute a Dry-Run Agent Invocation. Launch Claude Code or Google Antigravity pointing to the restored directory and invoke the skill against a sample codebase.
5. Step 5: Assert Expected Output Shape. Confirm that the agent follows the specified rubric, runs helper scripts without missing dependency errors, and produces deterministic results.

Git Sync for Technical Skill Libraries

Git is the cleanest manual option when the people maintaining the skills already use code review. It gives history, diffs, branches, and rollback. It also makes project skills easy to keep beside the repo conventions they describe.

mkdir -p ~/ai-skills
cd ~/ai-skills
git init
cp -R ~/.claude/skills/* .
git add .
git commit -m "Back up Claude skills"

On another machine, clone the repo and copy or link the selected folders into the Claude Code skills location.

git clone git@github.com:your-org/ai-skills.git ~/ai-skills
mkdir -p ~/.claude/skills
cp -R ~/ai-skills/diff-risk-review ~/.claude/skills/diff-risk-review

Git is strongest when it is the source of truth. It becomes messy when people edit local copies and forget to commit or pull. Pick one canonical library, then treat local Claude folders as synced outputs.

Managed Skill Library for Multi-Tool Sync

Git and cloud drives can move folders, but they do not solve the larger portability problem: your AI setup now spans Claude Code, Codex, ChatGPT, Claude, and MCP-connected agents. The workflow should be available where the work happens, not locked to whichever folder you remembered to copy.

Prompttly is a skill manager for AI agents — one library for your skills and prompts that syncs into Claude Code, Codex, ChatGPT, and Claude and is one hotkey away on your Mac, so your setup follows you across every machine, repo, and tool.

That means the library is the source of truth. The local folders in Claude Code and Codex are synced working copies. If you improve a skill on your desktop, the same skill can be kept available on a laptop instead of becoming another forgotten local file.

Portability is a workflow property, not a storage location.

A Safe Sync Process for Claude Skills

Use this process whenever you move from scattered local skills to a reliable synced library.

1. Inventory the current folders: List every skill in ~/.claude/skills and any project-level .claude/skills folders you rely on.
2. Pick the source of truth: Choose one library, repo, or skill manager where the canonical version lives.
3. Deduplicate similar skills: Merge overlapping folders before syncing. Do not keep code-review, review-code, and diff-review unless they truly do different jobs.
4. Preserve supporting files: Move each entire skill folder, not only the Markdown entrypoint.
5. Sync into each machine: Put working copies in the correct personal or project folder for each computer.
6. Test with the same task: Run one realistic prompt on both machines and compare whether the process and output match.

The test step matters. A folder can exist on both machines and still be wrong if the description changed, a template is missing, or Claude selects a different skill.

Before and After: From Scattered Files to Portable Skills

If the problem is team distribution rather than just multi-machine sync, use the Claude Skills team sharing guide to define owners, review updates, and roll out approved skills without private forks.

Common Mistakes When Syncing and Backing Up Skills

• Syncing only the entrypoint: Multi-file skills often depend on examples, templates, and references in subdirectories.
• Letting every machine be editable: Without a single source of truth, the newest version becomes a guessing game.
• Storing sensitive credentials in skills: Never hardcode API keys or database tokens inside skill scripts; use environment variables or MCP servers instead.
• Mixing personal and project scope: Repo-specific commands should not follow you into every unrelated workspace.
• Keeping duplicate names: Similar folder names make selection and maintenance harder.
• Skipping version history: If a useful skill gets worse, you need a way to recover the previous behavior.

If a skill matters enough to copy across computers, it matters enough to version.

The same rule applies across repositories. If several repos contain near-identical skill folders, use the Claude Skills across projects guide to turn the shared workflow into one global skill and keep repo facts local.

Post-Sync Verification Checklist

1. Confirm each skill folder exists on the target machine.
2. Confirm the entrypoint is named SKILL.md.
3. Open the skill and verify the frontmatter still matches the intended use case.
4. Check that templates, examples, references, and scripts traveled with the folder.
5. Run one real task that should trigger the skill.
6. Compare output structure across machines.
7. Remove stale duplicates after the synced version works.

For a deeper authoring, install, and invocation workflow, use how to install and use Claude Skills alongside our agent skills and SKILL.md guide.

Related Resources

Start from the resources hub for the full Claude Skills series. If you also need to synchronize Cursor rules, Codex instructions, and prompt libraries alongside your terminal skills, read our guide on how to sync prompts across laptops. Read our guide on Claude skills across projects and directory scopes to choose personal or project scope, use the new-computer transfer guide for one-time migrations, then use the usage guide to test behavior after sync. You can draft a reusable workflow with the free Claude Skill Creator.

People Also Ask About Syncing and Backing Up Claude Skills