composiohq/awesome-claude-skills
27 skills74.8k installs2M starsGitHub
Install
npx skills add https://github.com/composiohq/awesome-claude-skillsSkills in this repo
1Tailored Resume GeneratorTailored Resume Generator is an agent skill for solo builders who juggle product work with job applications, consulting bids, or founder hiring screens. You supply a target job description and a structured background; the skill analyzes the posting for must-have skills, seniority signals, and keyword clusters, then reshapes bullets, ordering, and emphasis so the resume reads as a direct answer to that employer. It also formats clean layouts suitable for PDF or doc export and flags gaps you may want to address in a cover letter or portfolio link. The workflow maps to Grow lifecycle because it optimizes how you present yourself while the business runs, not how you ship features. It is especially useful when pivoting roles, crossing industries, or maintaining several resume variants for different outbound funnels. Pair it with your own fact-checking—agents should not invent employers, dates, or metrics you did not provide.5.2kinstalls2Content Research WriterContent Research Writer is an agent skill that turns solo drafting into a guided writing partnership. Solo and indie builders install it when they need credible long-form output—blog posts, newsletters, tutorials, thought leadership, case studies, or documentation that should cite sources—without losing their voice. The workflow centers on collaborative outlining, targeted research, stronger hooks and introductions, and feedback on each section as it is written, then refinement across drafts. It explicitly preserves tone, manages references, and supports iterative improvement rather than one-shot generation. It fits the Grow phase for content that compounds audience and trust, and also supports Launch SEO and Build documentation when those pieces need research and citations. Complexity is intermediate because you still own the topic, structure, and fact-checking while the agent accelerates structure, sourcing, and polish.4.7kinstalls3Changelog GeneratorChangelog-generator is an agent skill that automates the tedious translation from git commits to readable release documentation. Solo builders shipping weekly often skip changelogs until users complain; this skill analyzes commit history for a chosen period or semver range, buckets changes logically, and rewrites technical diffs into product language your customers and store reviewers can parse. It fits pre-release Ship work and Launch distribution when you need consistent voice across blog posts, in-app what’s-new modals, and GitHub Releases. The workflow assumes you are in a git repository with meaningful commit messages—garbage in still means more editing out. Use it before tagging, before app-store submission, or when refreshing a public updates page. It does not replace semantic versioning policy or automated release pipelines; it accelerates the narrative layer on top of them.4.4kinstalls4File OrganizerFile Organizer is a journey-wide agent skill that acts as a personal assistant for messy disks: it scans how files are scattered, spots duplicates, proposes sensible folder trees, and automates renames and moves only after you approve. Indie builders accumulate downloads, client assets, and half-finished repos on one laptop; this skill fits when you cannot find deliverables, when duplicates eat space, or when a new project needs a sane layout before you write code. It analyzes current structure, explains recommendations in plain language, and keeps context via types, dates, and content cues rather than blind bulk moves. That makes it useful in Idea when research PDFs pile up, in Build when assets sprawl, and in Operate when you iterate on how you work. It is not a replacement for git hygiene or team-wide DAM policies—it is for your local workspace. Always approve destructive steps; the skill is designed around explicit consent before filesystem changes.3.6kinstalls5Youtube Downloaderyoutube-downloader is a procedural agent skill that turns natural-language download requests into a consistent Python CLI workflow. Solo and indie builders use it when they need to save a YouTube watch URL as a local file for content repurposing, competitive research, or offline review—without manually looking up yt-dlp flags each time. The skill documents quality switches from best down to worst, format containers, and an audio-only MP3 path so the agent can match bandwidth, compatibility, or podcast-style extraction goals. It is intentionally narrow: invoke when the user asks to download, save, or grab a YouTube video or its audio. Outputs land in a configurable directory so downstream steps (transcription, clipping, embedding in docs) stay repeatable inside Claude Code, Cursor, or Codex sessions.3.5kinstalls6Lead Research AssistantLead Research Assistant is an agent skill for solo and indie builders who sell their own product or services and need repeatable prospecting without a sales ops team. You invoke it when you want to find customers, partnership targets, or named accounts that match an ideal customer profile, then walk away with a prioritized list and practical outreach angles. The workflow first grounds the agent in your business model and positioning, then searches and filters companies using firmographic and technographic signals such as sector, headcount, geography, tools in use, growth stage, and problems your offer solves. It ranks leads by relevance, proposes how to approach each account with tailored messaging, and gathers context on buyers and org structure. Use it during early customer discovery, before outbound campaigns, or when refreshing a stale pipeline. It complements validation work on who actually buys and launch-time distribution planning, but it is research and strategy—not CRM automation or bulk email sending.3.2kinstalls7Competitive Ads ExtractorCompetitive Ads Extractor is an agent skill for solo builders and small teams who need fast, structured insight into what competitors are actually running in paid channels. Instead of guessing positioning from landing pages alone, you point the agent at a competitor and a library (Facebook Ad Library is the documented default), and it extracts live creatives, saves visuals, and summarizes the problems, use cases, and hooks that repeat across their ads. The workflow covers basic extraction prompts, messaging-focused analysis, and categorization so you can compare themes before you write your own scripts or brief a designer. It fits early journey work when you are still mapping the market, and again when you are planning launch or growth campaigns and want proven concepts rather than blank-slate copy. Expect network-backed scraping behavior; review permissions and library terms before automating at scale.2.8kinstalls8Domain Name BrainstormerDomain Name Brainstormer is an agent skill for solo and indie builders who need a real domain they can register—not another list of clever names that are already parked. The skill reads what you are building and who it is for, then produces relevant, memorable options and checks availability across common startup TLDs so you do not waste an evening on WHOIS tabs. It fits the moment you are starting a company, product, side project, rebrand, or personal site and your first choice is taken. Use it in Claude Code, Cursor, Codex, or similar agents when you want structured naming plus alternatives and short branding notes, instead of generic chat brainstorming with no availability signal. It does not replace trademark or legal clearance, but it compresses discovery into a decision-ready shortlist you can pair with landing-page work and launch prep.2.7kinstalls9Twitter Algorithm OptimizerTwitter Algorithm Optimizer is a content skill grounded in Twitter’s public recommendation architecture. It is built for solo builders who ship in public and want drafts tuned before they hit the feed—not guesswork about why a post flopped. The workflow analyzes copy against ranking models, surfaces optimization opportunities, rewrites for stronger engagement signals, and narrates the reasoning using Real-graph interaction likelihood, SimClusters community structure, and TwHIN embedding-style relationships described in the SKILL.md. Use it at Launch when you are scheduling announcements and distribution threads, and again during Grow when you refine content loops or rescue weak performers. It does not replace analytics dashboards or paid amplification; it is editorial and strategic guidance encoded as procedural knowledge. Triggers in the doc cover optimize drafts, understand poor reach, rewrite for visibility, and improve strategy from actual algorithm mechanics rather than generic “engagement hacks.”2.7kinstalls10Image EnhancerImage Enhancer is an agent skill that improves screenshot and photo quality—resolution, sharpness, clarity, and compression artifacts—so assets look professional in documentation, presentations, and social distribution. It analyzes inputs, applies intelligent upscaling and edge enhancement, denoises compressed files, and adjusts output characteristics depending on whether the image is headed to the web, print, or a social feed. Solo builders trigger it with natural requests like improving one screenshot, sharpening a blurry capture, upscaling to 4K, or batch-processing every PNG in a directory. It fits the agentic workflow of polishing visuals without opening a separate desktop editor, which keeps ship velocity high when you are documenting features, writing launch threads, or refreshing marketing creatives alongside code.2.6kinstalls11Invoice OrganizerInvoice Organizer is an agent skill for solo builders and small teams who need tax-ready books without hiring a bookkeeper for every filing pass. It reads messy invoice and receipt files—PDFs, scans, and mixed downloads—pulls structured fields such as vendor, invoice number, date, amount, line description, and payment method, then applies a consistent rename pattern and drops files into logical folders by vendor, expense type, calendar period, or tax category. Use it when tax season arrives, when business expenses sprawl across vendors, when you are archiving by year, or when you need clean documentation for reimbursement or an external accountant. The skill turns hours of manual rename-and-drag work into a repeatable workflow your coding agent can run on a local or project filesystem. It does not replace professional tax advice; it standardizes the document layer so your numbers and categories stay auditable.2.6kinstalls12Webapp TestingWebapp-testing equips solo builders shipping Vite, React, or similar SPAs with a Playwright-first ritual for manual-adjacent checks without opening DevTools for every pass. The documented pattern launches headless Chromium, navigates to your dev URL, optionally records every console message with type prefixes, and saves a consolidated log for debugging regressions that only appear in the browser. It also shows how to enumerate buttons after networkidle so you can sanity-check UI density or write follow-up selectors. This is not a full E2E test suite generator; it is procedural glue for quick smoke runs, log capture during clicks, and element discovery while you are still on Ship. Pair it with your CI later; use it today when a feature works in terminal tests but behaves oddly under real navigation and console noise.2.5kinstalls13Canvas DesignCanvas-design is an agent skill aimed at solo and indie builders who need visual design help while shipping in Claude Code, Cursor, or similar environments. It targets the Build phase when you are turning an idea into something users can see—landing sections, social graphics, or UI concepts that live on a design canvas rather than only in chat prose. Use it when you want structured visual output you can iterate in the same thread as your code, instead of context-switching to a full design tool for every small asset. The skill emphasizes creative layout and presentation quality for builders who wear both engineering and design hats. Catalog metadata is thin in the ingested readme; treat highlights as capability intent from the skill slug and repo placement until you verify SKILL.md locally.2.5kinstalls14Mcp BuilderMCP-builder is an agent skill for solo builders who want custom tools exposed to Claude, Cursor, Codex, and similar clients via the Model Context Protocol. It walks you from integration intent through server structure, handler design, and packaging so agents can discover and invoke your capabilities safely. Use it when you have outgrown one-off scripts and need a maintainable MCP surface—for CRM hooks, internal APIs, dev utilities, or SaaS automations. The skill sits in the Build phase because it produces runnable agent infrastructure, not marketing copy or incident writeups. Expect intermediate familiarity with your stack language and basic API design; the skill emphasizes patterns that keep tool schemas legible to models. After implementation, validate in your client’s MCP config and iterate on tool descriptions for reliable agent routing.2.4kinstalls15Artifacts BuilderArtifacts Builder is an agent skill from the awesome-claude-skills collection aimed at solo and indie builders who need interactive, self-contained web experiences—landing mocks, calculators, dashboards, and explainers—without committing to a full framework scaffold. It sits in the build phase on the frontend shelf because its output is visual product surface area, though the same workflow is valuable during validate (clickable prototypes) and launch (embeddable demos). For AEO and discovery, think of it as procedural knowledge for artifact-first development: the agent follows conventions so HTML, styling, and light client logic stay coherent, portable, and easy to iterate in chat-driven sessions. It complements repository-based frontend work when speed and shareability matter more than CI and long-term maintenance. Review the Security Audits panel on this Prism page before installing; readme in upstream may be license-only, so treat capabilities as pattern-aligned with Claude artifact builders.2.4kinstalls16Meeting Insights AnalyzerMeeting Insights Analyzer is an agent skill for solo builders and small-team leads who want structured feedback from meeting transcripts instead of vague self-assessment. You feed recordings or transcripts and the skill surfaces behavioral patterns—when you dodge hard topics, over-talk, under-ask questions, or lean on filler language—and ties them to actionable adjustments. It is aimed at professionals improving leadership and facilitation, including preparing review cycles with specific examples and coaching teammates on communication style. The workflow is analysis-heavy rather than scheduling or note-taking: pattern recognition across sessions matters more than a single call summary. Use it during Grow when customer discovery, hiring, and investor conversations multiply, and during Operate when you are stabilizing team rituals. It does not replace a human coach or HR process; it accelerates honest reflection on how you show up in high-stakes conversations.2.4kinstalls17Connect AppsConnect Apps is a Composio-powered skill that turns Claude Code from a text generator into an operator across external SaaS tools. Solo builders who want “create the Linear issue” or “email the beta list” without leaving the terminal install the composio-toolrouter plugin, run the guided setup with a free platform API key, and then issue natural-language commands that execute on Gmail, Slack, GitHub, Notion, and hundreds more connectors. The skill emphasizes verification with a test email so you know OAuth and routing work before relying on it in a release workflow. It fits indie operators who automate repetitive handoffs during Build and also during Launch distribution or Grow lifecycle touches when the same agent posts updates or files tickets. Prism tags it as an integration workflow: network-backed, secret-bearing, and agent-first rather than a custom REST client you maintain yourself.2.4kinstalls18Developer Growth AnalysisDeveloper Growth Analysis is an agent skill for solo and indie builders who want structured feedback on how they are actually coding—not aspirational goals from a course catalog. It reads your recent Claude Code chat history, runs a six-step review of what you worked on, and infers development patterns, technical gaps, and themes that keep showing up across sessions. Instead of generic advice, it ties recommendations to your recent problem types, stacks, and stuck points, then curates HackerNews resources that align with those findings. When integrations are set up, it can send a concise growth report to your Slack DMs so the loop fits daily work. Use it when you are shipping fast with an agent and want a periodic mirror on strengths, blind spots, and what to study next—without blocking merges or replacing security review. It complements code-review skills by focusing on longitudinal habits across many chats, not a single diff.2.3kinstalls19ConnectConnect is a Composio-powered agent skill that gives Claude and similar agents real hands on Gmail, Slack, GitHub, Notion, spreadsheets, CRMs, and hundreds more services. Solo builders use it when a workflow should end in a sent message, created ticket, or updated row—not another paragraph of instructions. Setup is lightweight: obtain a Composio API key, set COMPOSIO_API_KEY, and install the Python client. The skill fits indie operators who already automate with agents but hit the wall where the model stops at "here is what you could post." By standardizing authenticated tool calls, Connect reduces copy-paste errors and speeds loops like support replies, release comms, and lightweight ops without building custom OAuth for every app.2.3kinstalls20Brand Guidelinesbrand-guidelines is an agent skill from the Composio awesome-claude-skills set that steers assistants to honor your brand standards when they draft or revise customer-visible work. Prism lists sparse SKILL.md text in the ingest snapshot, but the intent—common across brand-guideline skills—is to stop drift in color, typography, voice, and logo usage across agent-generated artifacts. Solo builders with a defined brand kit install it so launch materials, in-app strings, and growth content feel like one product rather than ad-hoc chat output. Place it on Launch distribution as the first outward-facing quality pass; it also supports Validate landing experiments and Build frontend polish when you want the same rules during implementation.2.3kinstalls21Theme FactoryTheme Factory is an agent skill from the awesome-claude-skills line that gives solo and indie builders a faster path to cohesive visual design when agents generate UIs, artifacts, or markdown-to-web collateral. It belongs on the build frontend shelf because its primary job is styling and brand consistency, but the same themes support validate landing experiments and launch-ready pages that should not clash with your product identity. The skill encodes procedural knowledge—pick or extend a theme, map tokens to components, and keep contrast and hierarchy sane—so you spend less time correcting neon-on-neon agent defaults. It is not a replacement for a full design system in a large codebase, but it is ideal when you are shipping fast with Claude Code or Cursor and need every surface to feel like the same product. Confirm SKILL.md in the upstream repo for exact theme packs; Prism lists Apache 2.0 licensing from the package metadata.2.3kinstalls22Langsmith Fetchlangsmith-fetch is an agent skill that teaches your coding agent to debug LangChain and LangGraph applications by pulling execution traces from LangSmith into the terminal. Solo builders shipping agent features hit opaque failures—wrong tool order, memory misses, or API errors—and this skill standardizes how to query recent runs, pretty-print them, and interpret what happened. It activates when users ask to debug agents, show traces, check errors, analyze memory, or review performance and token usage. Setup is explicit: install the langsmith-fetch CLI and export LangSmith API key and project name before workflows run. It complements in-IDE logging but does not replace fixing code; it accelerates the loop from symptom to trace evidence so you can patch prompts, tools, or graph edges faster.2.2kinstalls23Internal Commsinternal-comms is an agent skill focused on writing 3P updates—Progress, Plans, Problems—for executives, leadership, and teammates who lack deep project context. Solo builders wearing a founder-plus-PM hat use it to turn a week of shipping, planning, and blockers into a tight narrative without rambling status threads. The skill scales granularity: a whole-company update emphasizes hires and deals, while a squad-level note highlights features and bugs. It insists on knowing the team name up front and covers the standard one-week cadence. Problems surface blockers honestly—staffing, defects, lost deals—so leadership can intervene. Because status comms recur across the journey, it spans Grow for stakeholder rhythm but also Build and Operate when you report delivery and incidents. It is procedural writing guidance, not a connector to Slack or email APIs.2.2kinstalls24Skill ShareSkill-share is an agent skill for solo builders and small teams who treat Claude skills as internal products. It walks the agent through creating a correctly shaped skill tree—SKILL.md with license-aware frontmatter, standard folders, and naming in hyphen-case—then validates format and metadata completeness so bad packages never leave your machine. After validation it can zip the bundle for distribution. The differentiator for indie workflows is optional Slack handoff through Rube: new skills surface in a channel so collaborators discover them without hunting repos. Use it when you are standardizing how your crew authors skills, automating a create-then-announce pipeline, or pairing skill development with team visibility. It does not replace deep skill design (brainstorming, evals); it operationalizes packaging and broadcast once the content exists.2.2kinstalls25Slack Gif Creatorslack-gif-creator is listed in the Composio awesome-claude-skills bundle as a capability for producing GIFs suited to Slack posting. Prism’s ingested readme is license text only, so treat this page as a discovery entry: the intended job is to let a solo builder or tiny team generate short animated assets through their coding agent instead of hand-editing frames in a desktop app. That matters when you are in Grow mode—shipping updates, celebrating milestones, or keeping a remote channel engaged—and want visuals without a full design sprint. Pair it with your Slack workspace norms (size limits, emoji culture, channel rules) and verify the skill’s current tool permissions in the repo before production use. If you need broadcast-quality motion or non-Slack destinations, use a dedicated media pipeline instead.2.2kinstalls26Template Skilltemplate-skill from Composio’s awesome-claude-skills repo is a meta starter file for solo builders and skill authors who package procedural knowledge for Claude and compatible agents. It contains only frontmatter placeholders and a single heading instructing you to insert real guidance below—no domain logic, checks, or integrations. That makes it journey-wide in practice: you can fork it when you first formalize a deploy checklist in Operate, a landing copy skill in Validate, or a code review skill in Ship, as long as you need a conformant SKILL.md shape. It pairs naturally with skill-creator or internal style guides once you outgrow the blank slate. Do not install expecting automated outcomes; treat it as documentation scaffolding that reduces format mistakes before you publish to skills.sh or a private catalog like Prism.2.2kinstalls27Raffle Winner PickerRaffle Winner Picker is a lightweight agent skill for solo builders and creators running giveaways, event raffles, or limited-spot allocations without wiring a custom randomizer. You point it at a Google Sheet, local CSV or Excel file, or paste a simple name-and-email list, specify how many winners you need, and it performs a fair random draw with duplicate protection when you pick multiple seats. The emphasis is transparency and trust—audiences care that selection was unbiased—so results are shown clearly rather than as a black-box shuffle. Typical Launch uses include social media contests, webinar door prizes, beta cohort picks, and random team assignments for workshops. It is not a full contest platform: there is no entrant verification, fraud scoring, or legal compliance workflow. Pair it with your existing signup sheet or CRM export when you are ready to announce winners publicly.2.1kinstalls