What are Lovable Skills?
Lovable Skills are markdown playbooks (one SKILL.md per Skill) saved at workspace level. When your request matches a Skill's description, Lovable loads it on demand and follows the instructions — the same model as Anthropic's Claude Agent Skills.
Three required parts: name, description, instructions
Every Lovable Skill is a single SKILL.md file. Optional bundled files can travel with it (templates, reference docs, scripts) — up to 200 files, 10 MB total.
---
name: launch-checklist
description: Use when I say I'm about to launch,
ship, share, or release a project, or when I
ask whether it is ready to go live. Run the
checklist below before giving a go/no-go.
---
# Launch checklist
Walk through every item. Report pass / fail /
"needs manual check". Do not skip ahead.
## Account flows
- Sign-up, sign-in, sign-out and password reset
are wired end to end.
- Signed-out users cannot reach private routes.
...Name
Lowercase, hyphens only, up to 64 chars. Permanent.
Description
Start with 'Use when…'. This is how Lovable decides whether to load the Skill.
Instructions
The markdown playbook. Up to 100,000 characters. Short, opinionated, with examples.
Five official ways to add a Lovable Skill
Workspace owners and admins manage Skills from Settings → Skills → Add. Four routes live there; the fifth happens inside a project chat.
Build with Lovable
Guided chat: Lovable asks what the Skill should do, when it should trigger, and drafts the SKILL.md for you to approve.
Write manually
Three fields — name, description, content — and click Add skill. Available across the workspace immediately.
Import from GitHub
Paste a public GitHub URL (whole repo, subdirectory, or blob URL pointing at SKILL.md). Lovable downloads, validates and publishes it.
Upload a ZIP
Drag a .zip or .skill archive (up to 50 MB) containing a SKILL.md plus bundled files. Same shape as Claude exports — fully portable.
Save from a project chat
After Lovable nails a task, say "save that as a skill" and describe when it should apply. Lovable drafts it; you approve.
Two ways Lovable applies a Skill
Automatic
When your request matches the Skill's "Use when…" description, Lovable loads it without you doing anything. Just describe what you want.
Slash command
Type /skill-name in chat, then your prompt. The Skill is the how, your prompt is the what.
Lovable Skills vs Lovable Knowledge vs Lovable Prompts
| Dimension | Knowledge | Skill | Prompt |
|---|---|---|---|
| Loaded | Always | On demand when description matches | Only when you type it |
| Scope | Whole workspace, every message | Whole workspace, matching tasks | One chat |
| Best for | Standards, brand voice, glossary | Repeatable playbooks (launch checklist, Stripe setup) | Ad-hoc requests |
| Storage | Workspace knowledge | Workspace Skills (SKILL.md) | Chat message |
Rule of thumb: if it applies to every message, it's workspace knowledge. If it only matters for specific kinds of tasks, it's a Skill. If you'll only ever say it once, it's a Lovable prompt.
Lovable Skills FAQ
What is a Lovable Skill?
A Lovable Skill is a markdown playbook (SKILL.md) saved at workspace level. It has a name, a 'Use when…' description, and instructions Lovable follows when the description matches your request. Skills are loaded on demand, so you can keep many focused Skills without slowing the agent down.
How are Lovable Skills different from Lovable knowledge?
Knowledge is always in context — use it for rules that apply to every message (coding standards, brand voice). Lovable Skills are loaded only when the request matches — use them for task-specific playbooks (launch checklist, Stripe setup, SEO review).
How do I add a Lovable Skill?
Five official ways: Build with Lovable (guided chat), Write manually, Import from GitHub URL, Upload a .zip, or 'save that as a skill' from inside a project chat. All four manage from Settings → Skills.
Can Lovable Skills be shared with my team?
Yes — every Skill defined in a workspace is automatically available to every project in that workspace. Workspace owners and admins control creation, editing and automatic-use toggles.
Are Lovable Skills the same as Claude Skills?
They use the same SKILL.md shape as Anthropic's Claude Agent Skills, so you can move a Skill between Lovable and Claude as a .zip with no rewrites.
What are the limits?
Each SKILL.md can be up to 100,000 characters. A Skill can include up to 200 bundled files (1 MB each, 10 MB total). Uploaded archives can be up to 50 MB.
Skip writing your own. Buy battle-tested Lovable Skills.
40+ production Skills for Stripe, Supabase RLS, SEO and AI chat — written by a Diamond Certified Lovable expert. From £29, one-time. Upload ZIP straight into Lovable.
Browse the Lovable Skill packs