Recommended Free Tools
For reusable Claude Code workflows, use a skill when you need structured instructions, optional supporting files, or the possibility of automatic selection. Use a custom command file for a lightweight, explicitly named slash command—especially if your project already uses that convention. For unattended scripts, use claude -p; for jobs triggered by GitHub events, configure Claude Code GitHub Actions. These are different layers of automation, not interchangeable features.
How Claude Code skills and custom commands differ
A skill is a directory built around a SKILL.md file. That file combines YAML frontmatter, which describes the skill and its behavior, with Markdown instructions Claude follows when the skill runs. A skill can also include supporting files. Its name, usually the directory name, becomes the slash-command name; its description helps Claude decide whether to load it for a task.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
As an Amazon Associate I earn from qualifying purchases.
Skills can be invoked directly, such as /summarize-changes, or selected automatically when Claude judges the description relevant. Automatic selection is not a guarantee for every vaguely related request: the description and invocation settings shape whether Claude uses a skill. Anthropic’s skills reference describes the frontmatter and invocation controls, including disable-model-invocation, which can prevent automatic model invocation when predictable, user-directed use is important.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →| Mechanism | Best fit | How it starts | What it can contain |
|---|---|---|---|
Skill in .claude/skills/ or ~/.claude/skills/ |
Repeatable instructions, team workflows, or a workflow that needs supporting resources | Slash command or eligible automatic selection | SKILL.md plus supporting files |
Custom command file in .claude/commands/ |
A lightweight, explicitly named slash-command convention | User invokes the command | Command file; less structure than a skill |
claude -p |
Scripts and non-interactive runs | CLI invocation | A prompt and deliberately configured tool permissions |
| Claude Code GitHub Actions | Repository automation tied to GitHub activity or a supplied prompt | Configured GitHub workflow | Workflow configuration, credentials or an eligible token/federation method, and repository permissions |
Claude Code continues to support files in .claude/commands/, but skills provide frontmatter and a directory for resources. If a skill and a command file use the same name, the skill takes precedence. Choose based on scope, triggering, complexity, and where the workflow runs—not on an assumption that one mechanism is always best.
#1 Best Overall
Where to put a skill and who can use it
- Personal: Put a skill under
~/.claude/skills/<skill-name>/to make it available as a user-level workflow. - Project: Put a shared skill under
.claude/skills/<skill-name>/and commit it so teammates can use it with the repository. - Organization or additional locations: Claude Code documents enterprise and additional-directory skill locations in its skills reference.
Discovery depends on context. A skill in a nested project directory may not appear when a session starts above that directory; Claude may discover it after entering the directory or after the user adds it. Monorepos, worktrees, trust, and Claude Code version can also affect what is visible. Check the current documentation and the version you have installed before relying on version-specific discovery behavior.
Review and trust shared project configuration as appropriate before using it. Claude Code distinguishes committed project settings in .claude/settings.json from personal project-level overrides in .claude/settings.local.json; the latter is normally kept out of commits. See Anthropic’s settings files and precedence reference for current scope and precedence details.
How to create a reusable skill
The following setup creates a personal skill. To share it with a repository instead, use .claude/skills/ as the parent directory and commit the skill after reviewing its contents.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problems- Create a directory named for the workflow:
mkdir -p ~/.claude/skills/summarize-changes. - Create
~/.claude/skills/summarize-changes/SKILL.md. Start the file with YAML frontmatter between---markers. Give it a useful description that says what work it handles and when it applies; follow the frontmatter format in the official skills reference. - Below the frontmatter, write the actual Markdown instructions: the inputs Claude should inspect, the output you expect, and any constraints that make the result useful. Add supporting files in the skill directory if the workflow needs them.
- Start or continue Claude Code in a context where the skill is discoverable, then invoke it as
/summarize-changes. If you need the skill to run only when explicitly requested, configure invocation behavior using the documented frontmatter controls rather than relying on the description alone.
Keep instructions specific enough to produce a repeatable result without assuming Claude will invoke the skill for every loosely related task. For team use, review any shared files and settings before committing them.
How to create a custom slash command
For a simple named command, create a Markdown command file in .claude/commands/ and invoke it by its filename as a slash command. This remains a supported convention, but it is the lighter-weight choice: skills add frontmatter, automatic-selection behavior, and a directory for related resources. If you want those capabilities, create a skill instead. Avoid giving a command file the same name as a skill, because the skill takes precedence.
How to automate Claude Code in scripts and CI
Run a non-interactive CLI prompt
Use claude -p "..." when a script needs to send a prompt without opening an interactive session. Anthropic’s CLI reference identifies print mode as useful for scripting and automation. Configure tool access deliberately—for example, with CLI options such as --allowedTools—so the run has only the permissions the job is expected to need. A skill or slash command by itself does not configure a non-interactive job or its permissions.
Rank #4
Run from a GitHub event or supplied prompt
For GitHub-triggered automation, use Claude Code GitHub Actions. The documented integration supports mention-triggered operation and prompt-supplied automation. Setup requires a workflow configuration, a suitable credential method (or eligible token/federation method), and appropriate repository permissions. Follow the current official setup and security guidance for the credential and permissions model you choose; a slash command alone does not install or authorize a GitHub integration.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Choose the workflow that matches the job
- Use a personal skill when you want a reusable workflow for your own sessions.
- Use a project skill when a team should share structured instructions and supporting assets in a repository.
- Use a command file when a small, explicitly invoked slash command is enough and the project uses that simple convention.
- Use
claude -pfor non-interactive terminal scripts, with tool permissions set for the job. - Use GitHub Actions when execution should be tied to GitHub activity or a configured prompt-driven workflow.
Claude Code’s documentation changes frequently. The official references for skills, CLI usage, GitHub Actions, and settings were checked on October 3, 2026; verify the guidance against your installed version and current documentation before relying on version-specific behavior.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
