Can you save prompts in Claude?
Claude has no saved-prompts list, but it gives you four places to keep instructions you reuse: Instructions for Claude (every chat), project instructions (every chat in one project), skills (named workflows you switch on, which is also where custom styles are moving) and, in Claude Code, skills and slash commands you run by name.
Those features hold standing guidance. If you want a set of separate prompts to pick from on demand, keep them outside the app in a file or prompt library and paste them in. This guide maps each option, with steps, and shows when each one fits.
Where does Claude keep instructions you reuse?
| Option | Where it applies | Plans (as of October 2026) | Best for |
|---|---|---|---|
| Instructions for Claude | Every chat | No plan limit listed | Your role, tone and format preferences |
| Project instructions | Chats in one project | Projects on all plans (Free: up to 5); one help article lists project instructions as paid plans only | Context for one client, product or codebase |
| Skills | When a request matches, or invoked by name | Free, Pro, Max, Team, Enterprise (needs code execution) | Repeatable workflows; former custom styles |
| Memory | Context from past chats | On by default for Free, Pro, Max | Facts about you and your work |
| Claude Code skills | Sessions on your machine or in a repo | Claude Code | Prompts you run with /name while coding |
| Outside library | Wherever you paste | Any | Many separate prompts with blanks |
Sources: Anthropic's help articles on personalization, projects, skills and memory.
How do you save a prompt as Instructions for Claude?
Use this for guidance you want in every chat. Anthropic's personalization article says "Any instructions you add here will be applied to all of your conversations with Claude."
- Click your initials in the lower-left corner and open Settings.
- Go to Your Claude (shown as Account in the older layout).
- Add the part of your prompt that should always apply, such as your role, preferred format and level of detail.
- Start a new chat to check the effect.
Keep it general. A task prompt in this field, such as a full code review checklist, will shape chats where you did not want it.
How do you save a prompt as project instructions?
A project gives a group of chats shared instructions and files. The same article says project instructions "only apply to chats within that project." Per Anthropic's projects guide:
- Click Projects in the sidebar, or go to claude.ai/projects.
- Click + New Project and give it a name and description (Claude can't see these).
- Click Set project instructions, paste your standing prompt, and click Save instructions.
- Add reference files to the project knowledge. "Anything you upload to this space will be used across all of your chats within that project."
Free users can create up to five projects, though Anthropic's personalization article lists project instructions as "paid plans only," so check your plan if the option is missing. As with any built-in option, the instructions stay in that project, so a good set gets copied by hand into the next. The Claude Projects guide covers knowledge limits and sharing.
What happened to Claude styles, and how do skills replace them?
Styles are moving to skills. Anthropic's help center article on the change said:
- Custom styles are migrated to skills automatically and appear in your account as skills.
- Migrated skills start disabled. Turn on the ones you want under Customize, then Skills.
- A migrated style runs as a slash command in the format
/{style-name}-style: the style name in lowercase, spaces turned into hyphens, punctuation removed, plus-style. For example, a style called "Client Email" would become/client-email-style. - The Concise, Explanatory and Formal default styles go away. Learning is kept as a skill.
- The Styles menu is being removed. Until the migration finishes, existing styles stay next to the new skills.
If you saved prompts as custom styles, check your Skills page and switch on the ones you still use.
How do you save a prompt as a skill in Claude?
A skill is a folder with a SKILL.md file: a name, a description of when to use it, and the instructions. Anthropic's skills help article says skills are available on Free, Pro, Max, Team and Enterprise plans, and "This feature requires code execution to be enabled."
- On Free, Pro or Max, go to Settings, then Capabilities, and turn on code execution and file creation. On Team and Enterprise, an owner enables code execution and Skills in organization settings.
- Write the skill: a folder whose name matches the skill name, with a
SKILL.mdinside. - Zip the folder.
- Go to Customize, then Skills, click +, choose + Create skill, then Upload a skill, and upload the zip.
- Toggle the skill on.
"Claude will automatically use these tools when relevant," so you don't have to call a skill yourself. If Claude ignores one, the article suggests checking that it is on and that its description clearly says when to use it. The Claude skills guide goes deeper, and the free SKILL.md generator turns a prompt into a valid SKILL.md in your browser with no sign-up.
Does Claude memory save prompts?
No. Anthropic's memory article says "Claude saves memory as a set of individual topics as you chat," and "Memory is on by default for Free, Pro, and Max plans." You can read, edit or delete topics under Settings, then Your Claude. Memory is for context such as your role and preferences, not for storing exact prompt text. See Claude memory for details.
How do you save prompts in Claude Code?
As skills. The Claude Code skills docs say "Custom commands have been merged into skills." A file at .claude/commands/deploy.md and a skill at .claude/skills/deploy/SKILL.md both create /deploy, and existing command files keep working.
- Personal:
~/.claude/skills/<skill-name>/SKILL.md, for all your projects on that machine. - Project:
.claude/skills/<skill-name>/SKILL.md, committed so your team gets it too.
A saved prompt that takes input looks like this:
---
name: fix-issue
description: Fix a GitHub issue
disable-model-invocation: true
---
Fix GitHub issue $ARGUMENTS following our coding standards.
Running /fix-issue 123 passes 123 into $ARGUMENTS. Setting disable-model-invocation: true means Claude can't run the skill on its own, which suits prompts with side effects. The Claude Code custom commands guide covers arguments and frontmatter in full.
When should a prompt stay a prompt instead of becoming a skill?
Turn a prompt into a skill when it is a procedure you repeat: several steps, a fixed output, maybe a template or script, and you want Claude to apply it without being asked. Keep it a prompt when it is short, used now and then, or changed every time you use it. Skills vs prompts covers the trade-off.
How do you keep one prompt library that works across Claude and other tools?
Built-in options keep prompts inside one Claude account and, for projects, inside one project. If you also use ChatGPT, Gemini or a coding agent, keep the source of truth outside all of them:
- Save each prompt as a Markdown file with blanks, such as
{{audience:engineers}}, for the parts that change. - Tag prompts by job and by tool.
- Copy always-on guidance into Instructions for Claude or a project.
- Turn repeated procedures into skills, and keep the
SKILL.mdfolders in the same place.
Kipt does this with a private GitHub repo you own: prompts with {{name:default}} blanks, tags, Cmd+K search, and Agent Skills you can install into Claude Code with one npx skills add command. The AI prompt library guide compares other approaches, and the same split for other apps is in how to save prompts in ChatGPT and how to save prompts in Gemini.