Reviewed skills
Popular agent skills, reviewed
What each skill really does, when your agent will use it, the lines worth reading, what to watch out for, and a score out of 100. Read the review before you install.
- 30 skills
- 74 average
- 15 worth installing as they are
Best first
- A88/100Install it
docxanthropics/skills
“Anthropic's Word skill: a tight list of docx pitfalls, plus real scripts for redlines and comments”
Creates, reads and edits .docx files. New documents are built with the docx npm library, guided by a list of known pitfalls. Existing ones are edited as raw XML. Helper scripts handle run merging, validation, tracked changes, comments and rendering pages for a visual check.
- A87/100Install it
xlsxanthropics/skills
“Anthropic's xlsx skill: dense, battle-tested spreadsheet rules that catch the silent formula bugs”
Guides the agent through creating, editing and reading Excel and CSV files with openpyxl and pandas. It then recalculates every formula through LibreOffice to catch errors. Most of it is hard-won gotchas about formulas, cached values, links and finance formatting.
- A83/100Install it
verification-before-completionobra/superpowers
“A short, strict rule that makes your agent show real test output before it says "done." Worth having.”
A discipline skill: before the agent claims work is fixed, passing or complete, it must run the command that proves it, read the output, and report what that output shows. No commands of its own, just a gate.
- A81/100Install it
writing-plansobra/superpowers
“Turns a spec into a tight TDD plan of exact files, signatures and tests. Best with the rest of Superpowers.”
Before any code, it turns a spec into a saved implementation plan of small TDD tasks. Each task gives exact files, interfaces, tests and commands. It self-reviews against the spec, then asks you to review and pick how to run it.
- A81/100Install it
pptxanthropics/skills
“Anthropic's PPTX skill: a field manual of real pptxgenjs and OOXML gotchas, with a long design lecture attached”
Creates, edits and reads PowerPoint files. It uses pptxgenjs for new decks, unzip/edit/zip with helper scripts for templates, and markitdown for reading. It adds a design guide and a required validate-and-render QA loop.
- B78/100Install it
using-git-worktreesobra/superpowers
“Careful git worktree setup: checks first, asks first, then installs and tests. Mind the auto npm install.”
Before feature work, it checks whether the agent is already in a worktree, asks before creating one, and prefers the harness's own worktree tool over git. It then installs dependencies and runs the tests to confirm a clean baseline.
- B78/100Install it
pdfanthropics/skills
“Anthropic's PDF skill: a solid cookbook for every PDF chore, with a sharp trigger and few checks”
A reference of Python and CLI recipes for reading, merging, splitting, rotating, watermarking, encrypting, creating and OCRing PDFs. Form filling is handed off to forms.md, and advanced topics to reference.md.
- B78/100Install it
frontend-designanthropics/skills
“Anthropic's anti-template design guide: a sharp list of AI-UI clichés and a plan-first process, no scripts”
Pushes the agent to design frontends that don't look generated. It grounds choices in the subject, plans a small token system, checks it against a list of common AI-design defaults, then builds and self-critiques. It closes with UI copywriting rules.
- B77/100Install it
systematic-debuggingobra/superpowers
“Systematic debugging: a strict root-cause-first process that stops the agent guessing at fixes. Solid.”
A four-phase debugging discipline: investigate the root cause, compare against working code, test one hypothesis at a time, then fix with a failing test first. After three failed fixes, it makes the agent stop and question the design with you.
- B77/100Install it
doc-coauthoringanthropics/skills
“A patient doc-writing coach that tests your draft on a fresh Claude before your team ever sees it”
Walks the user through writing a spec, proposal or decision doc in three stages: gather context, build each section by brainstorming and curating, then test the draft on a context-free Claude to find blind spots.
- B76/100Worth a look
vercel-composition-patternsvercel-labs/agent-skills
“Tidy index of React composition rules; the real lessons live in rule files this review couldn't see”
A short index of eight React composition rules: avoid boolean props, use compound components, lift state into providers, and adopt React 19 APIs. Each rule is one line here. Explanations and code examples are said to be in rules/*.md and AGENTS.md.
- B76/100Worth a look
vercel-react-best-practicesvercel-labs/agent-skills
“Vercel's 70-rule React speed cheat sheet: a sharp index, but the actual advice lives in other files”
A prioritized index of 70 React and Next.js performance rules from Vercel, grouped into 8 categories from waterfalls to micro-optimizations. Each rule gets one line here. Explanations and code examples are in rules/*.md and AGENTS.md.
- B76/100Worth a look
requesting-code-reviewobra/superpowers
“A tidy recipe for handing your diff to a fresh-eyed reviewer subagent, but the template it needs is offstage”
Tells the agent to grab base and head commit SHAs, dispatch a general-purpose subagent with a filled-in review template, then triage what comes back: fix Critical now, Important before moving on, note Minor.
- B76/100Worth a look
web-artifacts-builderanthropics/skills
“Anthropic's React-to-single-HTML artifact kit: lean and useful, but cites a section that doesn't exist”
Scaffolds a React + TypeScript + Tailwind + shadcn/ui project with a script, has the agent edit it, then bundles everything into one self-contained bundle.html to show as a claude.ai artifact.
- B75/100Install it
skill-creatoranthropics/skills
“Anthropic's skill-maker: a real test loop with baselines and benchmarks, inside a chatty 486-line pep talk”
Guides the agent through building or improving a skill: interview the user, draft SKILL.md, run test prompts with and without the skill, grade and benchmark the runs, collect feedback in a browser viewer, and iterate. It can also tune the description for triggering and package the result.
- B74/100Worth a look
web-design-guidelinesvercel-labs/agent-skills
“A UI review skill whose rules live on a URL: tidy and tiny, but it obeys whatever GitHub serves that day”
Reviews UI code against Vercel's Web Interface Guidelines. Before each review it downloads the current rules from a GitHub file, checks the given files against them, and reports findings as file:line. The rules themselves are not in the skill.
- B74/100Install it
webapp-testinganthropics/skills
“Anthropic's Playwright skill: lean, practical web app testing that never says what 'passed' looks like”
Guides the agent to test local web apps by writing Python Playwright scripts. A bundled helper starts and stops dev servers around the script, and a short decision tree covers static HTML, servers not yet running, and live apps.
- B73/100Install it
dispatching-parallel-agentsobra/superpowers
“A tidy playbook for splitting unrelated bugs across parallel subagents. Practical, if a little repetitive.”
Teaches the agent to split independent problems, usually separate failing test files, into self-contained subagent tasks. It sends them all in one response so they run concurrently, then reviews, checks for conflicts and runs the full suite.
- B73/100Install it
mcp-builderanthropics/skills
“Anthropic's MCP server playbook: a clear route from API docs to tested tools and evals, a bit wordy”
Guides the agent through building an MCP server in TypeScript or Python. It covers research, picking an SDK version, building the tools, then testing and writing 10 read-only evaluation questions. Detailed patterns live in linked reference guides.
- B72/100Worth a look
internal-commsanthropics/skills
“A tidy 26-line router for internal memos. The real templates live in example files this review can't see.”
Helps the agent write internal communications such as 3P updates, newsletters, FAQs and status reports. It picks the type, loads a matching guideline from examples/, and follows that file's format and tone.
- B72/100Worth a look
brand-guidelinesanthropics/skills
“Anthropic's colors and fonts in one tidy card. Solid hex codes, but it describes a tool that doesn't ship”
Gives the agent Anthropic's brand palette (four neutrals, three accents) and fonts (Poppins for headings, Lora for body, with fallbacks). It also describes styling features as if a script applied them, but no script or steps are included.
- B72/100Install it
theme-factoryanthropics/skills
“Ten ready-made color and font themes for decks and docs, with a polite pick-one-first workflow”
Gives the agent 10 preset color and font themes, stored in a themes/ folder and previewed in a showcase PDF, to apply to slides, docs or HTML pages. It shows the options, waits for the user to choose, then applies the theme. It can also invent a new theme on request.
- B70/100Worth a look
test-driven-developmentobra/superpowers
“A strict TDD drill for agents: red, green, refactor, and delete any code that dared to come first”
Makes the agent work test-first: write one failing test, watch it fail for the right reason, write minimal code, run the whole suite, then refactor. Code written before its test gets deleted, and a long table rebuts the usual excuses.
- B69/100Worth a look
algorithmic-artanthropics/skills
“Solid p5.js generative-art skill that writes a manifesto first and tells itself it's a master, often”
Turns a request for code art into two outputs: a 4-6 paragraph "algorithmic philosophy" in markdown, then one self-contained p5.js HTML page. The page is built on a bundled Anthropic-branded template with seed navigation and parameter sliders.
- B68/100Worth a look
executing-plansobra/superpowers
“A strict, ledger-keeping way to run a coding plan yourself: TDD-gated, recoverable, a bit wordy”
Runs a written implementation plan task by task in the current session, with no per-task subagents. Each task is TDD-gated and logged in a ledger file that survives compaction. Plan defects become recorded rulings, and one fresh reviewer checks the whole branch at the end.
- B67/100Worth a look
brainstormingobra/superpowers
“A sharp 'find out what they want before you build' skill whose trigger fires on nearly everything”
Makes the agent pause before building. It reads the project, asks one question at a time, plays back what it heard, offers two or three approaches, and writes a design doc. Nothing gets built until the user approves it; then it hands off to a planning skill.
- C66/100Worth a look
writing-skillsobra/superpowers
“Skill-writing as TDD: make an agent fail first, then write the fix. Sharp method, strict, leans on unseen files”
Treats writing an agent skill like test-driven development. You run pressure scenarios on a subagent without the skill, record how it fails, write the smallest skill that fixes that, then retest and close loopholes. It adds wording micro-tests and a creation checklist.
- C65/100Worth a look
subagent-driven-developmentobra/superpowers
“A strict subagent foreman for running coding plans: smart rules, few check-ins, half its manual elsewhere”
Runs a written implementation plan by sending each task to a fresh subagent, then a reviewer for spec and quality. It loops fixes, logs every step and decision to a ledger file that survives context loss, and ends with a whole-branch review.
- C61/100Worth a look
canvas-designanthropics/skills
“Poster-making skill with real taste, no rendering tools, and a made-up quote from the user”
Turns a request for a poster or art piece into two files. First a short 'design movement' manifesto saved as .md, then a single-page PDF or PNG that expresses it, followed by a polishing pass. Most of the file is about aesthetics and craft. It never says how to draw anything.
- C53/100Worth a look
claude-apianthropics/skills
“Anthropic's own Claude API reference: precise and current, but a 28k-token tome that fires on any LLM task”
An official reference for writing code against the Claude API and Anthropic SDKs. It covers model IDs, prices, thinking, caching, tool use, Managed Agents and cloud clients. Bare subcommands (migrate, cost-optimize, build-eval and others) run guided workflows.
How the score works
Five things, out of 100
Trigger, Focus, Lean, Concrete and Safe, from Anthropic’s Extend Claude with skills and agentskills.io’s the Agent Skills standard. Kin reads SKILL.md only, not the scripts beside it, and a skill may have changed since it was reviewed.
Skills run while you look away.
Briefkin shows every Claude Code, Codex and Cursor session on one board: working, waiting on you, done. Answer them from your phone. Free in the open beta.