Skills.
Give your AI a better way to work.
A skill is a set of written instructions that teaches an AI how to do one job the way it should be done: review a pull request, plan a migration, write the release notes.
Install one here and it travels with your account into Claude, Claude Code, Cursor and every other client you sign in with.
Filter by category
70,704 results · page 638 of 2,357
- View details
maintain-verification-skillSkillAI & models
Lets your agent audit and update a project's verification skill and feature map, submitting corrections as one PR.
Ready to connect★ 2k
- View details
make-pr-easy-to-reviewSkillAI & models
A make claude skill that tidies pull requests: cleans commit history, improves descriptions, adds reviewer notes.
Ready to connect★ 2k
- View details
no-commentsSkillAI & models
Lets your agent spawn a subagent that fixes accepted code review findings and offers encodings for claimed constraints.
Ready to connect★ 2k
- View details
poteto-modeSkillAI & models
Lets your agent follow poteto's working style: concise answers, simple code, deliberate subagents, and verified work.
Ready to connect★ 2k
- View details
principle-attack-the-premiseSkillAI & models
Helps your agent question shared assumptions behind failed fixes instead of repeating them.
Ready to connect★ 2k
- View details
principle-boundary-disciplineSkillAI & models
Guides your agent to put validation at system edges like config and APIs and keep business logic in pure functions.
★ 2k
- View details
principle-build-the-leverSkillAI & models
Guides your agent to build a rerunnable script or tool for non-trivial work instead of doing edits by hand.
★ 2k
- View details
principle-encode-lessons-in-structureSkillAI & models
Helps your agent turn repeated instructions and corrections into automated checks instead of repeating text.
Ready to connect★ 2k
- View details
principle-exhaust-the-design-spaceSkillMedia
Guides your agent to build 2-3 competing prototypes of a design and compare them before choosing one.
Ready to connect★ 2k
- View details
principle-experience-firstSkillAI & models
Guides your agent to favor polished, user-friendly product choices over quick or rough implementations.
★ 2k
- View details
principle-fix-root-causesSkillAI & models
Guides your agent to debug by tracing symptoms to their root cause and fixing them there.
Ready to connect★ 2k
- View details
principle-foundational-thinkingSkillAI & models
Guides your agent to plan data structures and work order before writing code so the rest becomes simpler.
★ 2k
- View details
principle-guard-the-context-windowSkillFiles & storage
Keeps your agent's working memory from filling up by sending bulk reading tasks to subagents and keeping only summaries.
★ 2k
- View details
principle-laziness-protocolSkillAI & models
Lets your agent keep code changes minimal when refactoring, favoring deletion and simple fixes over new abstractions.
Ready to connect★ 2k
- View details
principle-make-operations-idempotentSkillAI & models
Guides your agent to design commands and processing steps that can safely run again after crashes or retries.
Ready to connect★ 2k
- View details
principle-migrate-callers-then-delete-legacy-apisSkillAI & models
Guides your agent to migrate code callers to a new API and delete the old one in one change instead of keeping shims.
Ready to connect★ 2k
- View details
principle-minimize-reader-loadSkillAI & models
Helps your agent simplify code that is hard to follow by counting indirection layers and collapsing needless wrappers.
★ 2k
- View details
principle-model-the-domainSkillFiles & storage
Guides your agent to replace scattered if/else logic with a data structure that models the problem domain.
★ 2k
- View details
principle-never-block-on-the-humanSkillAI & models
Lets your agent proceed with reversible work without asking permission, reserving confirmation for irreversible actions.
★ 2k
- View details
principle-outcome-oriented-executionSkillAI & models
Guides your agent through planned rewrites and migrations by prioritizing the final architecture over keeping every intermediate step working.
★ 2k
- View details
principle-prove-it-worksSkillProductivity
Makes your agent verify a finished task by testing the real result instead of just claiming it's done.
★ 2k
- View details
principle-redesign-from-first-principlesSkillMedia
Guides your agent to redesign an existing design from scratch when adding a new requirement instead of patching it on.
Ready to connect★ 2k
- View details
principle-separate-before-serializing-shared-stateSkillFiles & storage
Guides your agent to split shared files or state between concurrent writers instead of serializing access.
★ 2k
- View details
principle-sequence-verifiable-unitsSkillAI & models
Guides your agent to break multi-step work into small, checkable units and stack commits so each step proves itself.
★ 2k
- View details
principle-subtract-before-you-addSkillAI & models
Guides your agent to remove dead code and redundant pieces before adding or refactoring new features.
Ready to connect★ 2k
- View details
principle-test-behavior-not-implementationSkillMonitoring & ops
Helps your agent write tests that catch real defects by checking behavior instead of implementation details.
Ready to connect★ 2k
- View details
principle-type-system-disciplineSkillAI & models
Guides your agent to design types in statically-typed languages so invalid states cannot even compile.
★ 2k
- View details
setup-pstackSkillAI & models
Lets your agent configure pstack models by detecting available models and writing a per-role override sheet.
Ready to connect★ 2k
- View details
show-me-your-workSkillMonitoring & ops
Lets your agent keep a decision log of what it did and why during long-running or unattended work.
★ 2k
- View details
swarmSkillAI & models
Lets your agent run many parallel sub-workers on a task and combine their results into one report.
Ready to connect★ 2k
Popular skills lists
- Best Claude skills for awesome
- Best Claude skills for github
- Best Claude skills for anthropic
- Best Claude skills for grill me
- Best Claude skills for caveman
- Best Claude skills for humanizer
- Best Claude skills for remotion
- Best Claude skills for superpowers
- Best Claude skills for obsidian
- Best Claude skills for frontend design
- All best lists
What is a skill?
A skill is plain text, usually a SKILL.md file and the scripts it refers to, written for an AI rather than for a person. It carries the steps, the house rules and the examples a good answer needs, so you stop pasting the same briefing into every new chat.
68,574 of the 70,704 skills listed here can be served through Ahel today, and they come from public repositories. Each one has its own page with the instructions themselves on it, so you can read what a skill will tell your AI to do before you install it.
Install one and every AI you use gets it
Installing a skill adds it to your workspace and turns it on in the same step. Claude Code surfaces it as a slash command; any client can read the full instructions with the skill_read tool.
Nothing is copied into a project folder. The instructions are served from your account, so the same skill is there in every AI you connect, and turning it off removes it from all of them at once.