akillness/oh-my-skills
127 skills23.9k installs5.1k starsGitHub
Install
npx skills add https://github.com/akillness/oh-my-skillsSkills in this repo
1Graphifygraphify is an agent skill in the akillness/oh-my-skills collection that converts tabular or relational datasets into charts, graph views, and visualization-ready outputs for reports, dashboards, and stakeholder documentation. The skill guides coding agents to select appropriate chart types, map columns to axes, label series clearly, and export visuals suitable for README files, internal wiki pages, sprint demos, and executive summaries. Developers reach for graphify when SQL query results, CSV exports, JSON metrics blobs, or ORM query output need quick transformation into bar, line, scatter, or network graph presentations without manually opening a separate BI tool. graphify fits mid-build and pre-demo moments where feature teams must communicate experiment results, usage trends, or A/B outcomes alongside code changes. Install through the oh-my-skills repository using the standard skills CLI so Claude Code or Cursor sessions can generate reproducible chart artifacts from repository-local data files during agent-assisted documentation workflows.413installs2Data Analysisdata-analysis is version 2.0 of the akillness/oh-my-skills analytics skill for turning messy exports into decision-ready findings. It enforces a staged workflow: frame the decision question, profile data trust with a minimum checklist covering row counts, schema types, nulls, duplicates, and time coverage, then choose among four analysis lanes—spreadsheet-scale triage, SQL slicing, notebook or statistical analysis, or stakeholder-ready summary. The skill separates observation from interpretation, documents caveats, and supports retention, cohort, funnel, conversion, telemetry, and KPI explanations using Read, Grep, Glob, and Bash tools. It routes outward to looker-studio-bigquery, log-analysis, or codebase-search when the primary job is dashboards, incident logs, or repo navigation. Reach for data-analysis when a CSV export, warehouse query, or experiment table needs quality checks and concise evidence before scoping or building a product bet.406installs3Code Refactoringcode-refactoring is a routing-first cleanup skill in akillness/oh-my-skills that performs behavior-preserving structural hardening on legacy modules. It starts from the cleanup packet teams actually have—local cleanup, fragile legacy areas, repeated migrations, codemod plans, or cleanup-heavy diff shaping—and routes work through references/intake-packets-and-route-outs.md. The skill extracts functions, renames APIs, reduces coupling, and plans codemods while preserving observable behavior, with eval coverage including search-first blast-radius route-outs. It explicitly routes performance tuning to performance-optimization and correctness bugs to debugging rather than absorbing unrelated work. Developers reach for code-refactoring when diffs need structural cleanup, legacy modules need decomposition, or repeated migration patterns need codemod planning without risky behavior changes.389installs4Rtkrtk is an oh-my-skills installation and operations skill (version 1.0) for Rust Token Killer, a high-performance CLI proxy from the akillness/rtk fork of rtk-ai/rtk that compresses verbose shell output before it reaches AI agents. The skill verifies installations with rtk gain, repairs the common cargo install rtk name-collision with Rust Type Kit, supports Homebrew, install.sh, or Cargo install paths, and runs rtk init for Claude Code, Codex, Gemini CLI, Cursor, Copilot, Windsurf, Cline, and OpenCode. Developers reach for rtk when agent sessions waste context on git log, cargo test, or pytest output and need hook-based or direct rtk git status, rtk read, rtk grep, and rtk test wrappers.389installs5Remotion Video Productionremotion-video-production is an agent skill from akillness/oh-my-skills that guides developers through building videos as React code using the Remotion framework. The skill helps define compositions, timelines, props, animations, and render commands so promos, explainers, social clips, and release announcements can be regenerated from templates when data or copy changes. Developers reach for remotion-video-production when marketing or product teams need repeatable, version-controlled video output tied to a codebase rather than one-off Premiere or Final Cut projects. It covers the build-phase workflow of authoring Remotion components, previewing in the Remotion studio, and exporting MP4 or GIF assets from parameterized React templates.383installs6Technical Writingtechnical-writing is an agent skill from akillness/oh-my-skills that helps developers produce clear, accurate developer-facing documentation while implementing features. The skill guides drafting and refining README files, API reference sections, architecture notes, setup guides, and in-repo how-to documents so shipped code includes usable prose for the next engineer or API consumer. Developers reach for technical-writing when a feature is code-complete but docs are missing, outdated, or too terse for onboarding and integration. It supports the build-phase docs subphase and often continues into ship when release notes and changelog entries must match the final API surface.353installs7Research Paper Writingresearch-paper-writing is an oh-my-skills agent workflow for ML, CV, and NLP academic manuscripts and rebuttals. It supports drafting or rewriting Abstract, Introduction, Method, Experiments, and Conclusion sections, improving paragraph flow, checking claim-evidence alignment, and running pre-submission self-review from a reviewer mindset. The skill also covers figure and table support, reviewer response letters, and camera-ready revision passes. Developers explicitly request it in Claude Code, Codex, or Gemini prompts when transforming research notes into submission-ready prose rather than informal blog posts. It complements separate deep-research or LaTeX formatting skills when teams need end-to-end paper pipelines, but this package focuses on section-level writing quality and evidence discipline for computer-science publication venues.338installs8Web Accessibilityweb-accessibility is an oh-my-skills agent skill for web teams shipping accessible interfaces. It guides agents to audit pages against WCAG expectations, resolve keyboard traps that block navigation, correct insufficient color contrast, and add or repair ARIA labels before release or compliance review. Developers invoke it when accessibility is a gate for production deploys, enterprise procurement, or public-sector requirements. The skill bridges manual QA gaps by turning common a11y failures—focus management, semantic labels, and contrast ratios—into concrete HTML and CSS fixes inside the codebase rather than external spreadsheet audits alone.330installs9Video Productionvideo-production is a Claude Code skill for end-to-end video creation aimed at software products. It guides developers through planning narrative structure, writing scripts, organizing shoots, and editing finished clips suitable for documentation embeds, social promotion, and landing page heroes. The skill covers product demos, tutorial walkthroughs, and short launch announcements rather than cinematic production pipelines. Developers reach for video-production when a feature or release needs a visual explanation that screenshots alone cannot convey, or when onboarding docs benefit from screencast-style guidance. Because it spans pre-production through edit-ready output, it fits launch and growth workflows where video is a distribution artifact alongside written content.322installs10Design Systemdesign-system is an akillness/oh-my-skills agent skill for defining shared frontend UI rules before polishing individual screens. It triggers on design system, design tokens, visual language, component library foundations, UI governance, cross-product consistency, and primitive naming requests. The skill classifies problems into foundations, cross-surface alignment, primitive governance, or handoff modes, then produces a compact Design System Packet covering scope, color and semantic token rules, typography and spacing scales, motion and breakpoint policy, primitive naming, surface guidance for landing and dashboard UIs, accessibility baselines, and explicit route-outs to ui-component-patterns, responsive-design, and web-accessibility. Developers reach for design-system when UIs feel inconsistent across landing pages, dashboards, forms, and libraries—not when a single component API or layout fix is the core ask. Reference docs include scope-boundaries.md and design-system-packet-template.md. Rules-before-examples discipline prevents repeated one-off redesigns by locking governance before code generation begins.308installs11Responsive Designresponsive-design is a frontend skill from akillness/oh-my-skills that helps developers implement and audit responsive web layouts. The skill covers breakpoint strategy, fluid grid systems, viewport-aware typography, and touch-friendly interaction targets so interfaces adapt cleanly from mobile to desktop. Developers reach for responsive-design when shipping multi-device web apps, fixing layout regressions on narrow viewports, or reviewing CSS and component structure before release. It pairs naturally with design-system work and complements media-query debugging when Tailwind, CSS Grid, or Flexbox layouts break at specific widths.306installs12Presentation Builderpresentation-builder is an akillness/oh-my-skills agent skill that converts engineering notes, product specs, or research summaries into structured slide decks and speaker-ready talk outlines. Developers and tech leads invoke it when preparing conference demos, investor pitches, sprint reviews, or internal enablement sessions and need narrative structure before opening Keynote, Google Slides, or Reveal.js exporters. The skill organizes content into slide sequences with section headers, key talking points, and transition cues so presenters focus on delivery instead of blank-slide paralysis. Reach for presentation-builder when a feature spec or incident postmortem must become a stakeholder deck, or when onboarding materials need consistent slide flow across engineering and product teams. presentation-builder keeps documentation-oriented planning inside the oh-my-skills collection, producing reusable outlines teams can hand to designers for visual polish.294installs13Google Workspacegoogle-workspace is an Apache-2.0 skill (version 1.1.0) from akillness/oh-my-skills that plans Google Workspace automations before writing code. It classifies requests across nine product surfaces—Docs, Sheets, Slides, Drive, Gmail, Calendar, Forms, Chat, and Admin SDK—and chooses Apps Script for bound triggers, direct REST plus google-api-python-client for backend services, or Admin SDK for directory tasks. Auth guidance covers OAuth user flows, service accounts, and domain-wide delegation with narrow scopes. Workflow packets include Form→Sheet→Gmail intake, Sheet approval queues, and reporting flows with idempotency and quota checks. Three reference files detail auth routing, service selection, and cross-service patterns. Invoke when automating spreadsheets, shared drives, inbox workflows, or domain admin changes.285installs14Backend Testingbackend-testing is a testing workflow skill from akillness/oh-my-skills that guides developers through designing and executing backend test coverage across unit, integration, and contract test layers for APIs and services. The skill targets pre-release validation and post-refactor regression checks where service boundaries, HTTP contracts, and database interactions need explicit verification. Developers reach for backend-testing when shipping API changes, extracting microservices, or hardening services that lack structured test plans. Without detailed manifest constraints in the source, treat it as a procedural test-design assistant rather than a framework-specific test generator.280installs15User Guide Writinguser-guide-writing is version 2.1.0 documentation skill in akillness/oh-my-skills that anchors customer-facing help authoring for coding agents. The SKILL.md defines five primary modes—getting-started, tutorial, how-to, faq, and release-help-update—and an eight-step workflow from mode classification through maintenance sync checks. Developers reach for user-guide-writing when new users need first-success guides, features need task articles, or shipped UI changes make existing help-center steps stale. The skill enforces observable click-and-see instructions, early prerequisites, success checks, and explicit route-outs to technical-writing, api-documentation, and changelog-maintenance when requests drift. It references four support docs for mode boundaries, structures, workflow checklists, and maintenance signals, and outputs single pages, guide-plus-FAQ packets, or release refresh lists rather than mixed-purpose handbooks.265installs16Surveysurvey is an oh-my-skills agent skill (version 2.1.7) for pre-implementation landscape research across Claude Code, Codex, Gemini CLI, and OpenCode. The workflow freezes one research question, runs four parallel lanes—context, solutions, actual behavior, and alternatives or platform-map—and saves reusable markdown into `.survey/{slug}/` with fixed headings for triage, context, solutions, and platform-map when relevant. A bundled `validate_survey_artifacts.py` script checks the artifact contract before handoff to planning or execution skills. Developers reach for survey when comparing agent platforms, mapping workflows, or surveying competitors and workarounds where claims must stay source-backed. The skill explicitly stays in research mode and routes architecture or implementation outward only after artifacts validate.255installs17Task Planningtask-planning is an oh-my-skills workflow skill that converts ambiguous feature requests into structured execution plans with ordered tasks, milestones, and acceptance criteria. The skill gives agents and human developers a shared checklist before implementation begins, reducing rework from unclear scope. Developers reach for task-planning when a ticket or prompt is too vague to code directly, when an agent needs explicit steps, or when a feature needs milestone boundaries and testable done criteria. It fits early in a build cycle as a lightweight PM layer inside agent-driven development.253installs18Codebase SearchCodebase-search teaches agents disciplined repository exploration—scoped greps, symbol chasing, dependency tracing, and read budgeting—so changes target the right files in big or unfamiliar projects.243installs19Bmad GdsImplements the BMAD Game Development System so teams plan mechanics, levels, narrative beats, and production milestones with structured prompts, keeping design, engineering, and content aligned on playable game delivery.242installs20File Organizationfile-organization helps agents systematically sort, rename, and restructure project files using repeatable folder conventions and cleanup rules, reducing repo noise and making assets easier to find during long-running development and operations.240installs21Grill With DocsGrill-with-docs runs structured grilling sessions that challenge your plan against the existing domain model, sharpen terminology, and update CONTEXT.md and ADRs inline so scope and architecture decisions are explicit before build.239installs22Looker Studio BigqueryGuides teams to wire BigQuery tables into Looker Studio for post-launch analytics dashboards, defining metrics and visualizations that track funnels, retention, and revenue so growth and product leaders can monitor live performance.239installs23Api Designapi-design from akillness/oh-my-skills guides structured API design during backend development, covering endpoints, schemas, authentication, versioning, errors, and integration boundaries so services are consistent and implementable.237installs24Ui Component PatternsGuides developers through proven UI component patterns for modern web apps: defining prop contracts, loading and error states, compound components, and accessible markup so interfaces stay consistent as features scale.237installs25OmcThe omc skill is the core orchestrator for the akillness/oh-my-skills ecosystem, helping developers bootstrap, discover, and consistently apply Claude Code skills across projects so agent-assisted coding stays predictable and low-friction.236installs26Log AnalysisSupports analyzing application and infrastructure logs from staging and production to diagnose failures, correlate incidents, trace error patterns, and identify root causes so operators can restore service and prevent recurring outages.235installs27Vibe KanbanProvides kanban-style project management for vibe-coding workflows: define columns and WIP limits, slice backlog into agent-friendly tasks, track status across sessions, and keep build momentum visible without enterprise PM overhead.235installs28Api DocumentationStructures production-grade API documentation: endpoint tables, authentication flows, payload schemas, curl examples, and error references so backend teams ship integrator-ready docs that reduce support load and speed partner onboarding.233installs29Task EstimationHelps estimate engineering tasks by decomposing features into bounded units, applying effort ranges, noting dependencies, and aligning totals with sprint capacity so agent-assisted builds ship on realistic timelines.232installs30OmxOMX from oh-my-skills is agent orchestration for Claude Code: discover, select, and execute the right skill for each step so multi-skill build workflows stay composable, repeatable, and easier to maintain across repos.231installs31Database Schema DesignHelps agents design relational and document database schemas during scoping: entities, keys, relationships, indexes, constraints, and migration plans tuned to expected queries, growth, and reporting needs.229installs32Standup MeetingStandup-meeting skill from oh-my-skills helps teams run consistent daily standups by turning scattered updates into structured yesterday-today-blockers summaries, highlighting impediments, and preserving sprint context for PM and engineering workflows.228installs33StrixSecurity-focused oh-my-skills entry for Strix-style pre-ship audits: map trust boundaries in game or agent stacks, test auth and API endpoints, flag secret leaks, review third-party dependencies, and produce actionable fixes before launch approval.228installs34AutoresearchAutoresearch skill automates early discovery: agents run repeatable research loops, collect sources, summarize evidence, and output actionable briefs so teams decide what to validate or build with less manual searching.227installs35Code ReviewOffers a repeatable code-review playbook for evaluating pull requests: checking logic errors, naming and structure, missing tests, security issues, and release readiness so reviewers give consistent, actionable feedback before merge.227installs36Marketing AutomationGuides agents through designing and implementing marketing automation workflows including segmentation, triggered emails, CRM sync, and lifecycle campaigns for SaaS and ecommerce products.227installs37Plannotatorplannotator adds structured annotations to PRDs and agent execution plans in oh-my-skills, marking assumptions, risks, dependencies, and done definitions so builders validate scope before implementation begins.227installs38OpencontextThe opencontext skill from akillness/oh-my-skills helps developers wire up structured, retrievable context for AI coding agents. It focuses on selecting relevant files, summarizing state, and packaging instructions so assistants execute tasks with fewer wrong assumptions and less manual copy-paste of repository background.226installs39Workflow AutomationGuides designing and wiring workflow automation for software teams: identify repetitive rituals, encode them as agent or script pipelines, connect issue trackers and git flows, and iterate triggers until manual coordination drops without losing oversight.226installs40Authentication SetupGuides implementing application authentication including provider selection, OAuth flows, session or token management, secure login endpoints, and credential handling so SaaS apps and agents can verify user identity before granting access.225installs41Changelog MaintenanceMaintains structured CHANGELOG files for semver releases by curating commits into Added, Changed, Fixed, and Breaking sections. Ensures release notes stay accurate, scannable, and ready for tags, PRs, and user communications.225installs42Performance OptimizationProvides structured guidance for diagnosing and fixing performance bottlenecks across frontend rendering, network calls, and backend hot paths. It helps agents prioritize measurable wins before shipping features that must meet latency and efficiency targets.225installs43Playwriterplaywriter equips agents to author and execute Playwright-based end-to-end tests for web applications, turning observed UI flows into maintainable regression suites that catch breakage in navigation, forms, and critical paths before shipping.225installs44State ManagementTeaches agents how to implement robust frontend state management for interactive applications, covering local state, shared stores, async updates, and component coordination. It supports building maintainable UIs with clear data flow during feature development.224installs45Bmad Ideabmad-idea applies the BMAD methodology’s early ideation pass inside Claude Code, guiding founders and PMs from fuzzy intent to concrete product directions. It structures brainstorming, problem framing, and lightweight requirements so teams know what to validate next instead of jumping straight into responsive layouts or backend APIs.223installs46Security Best PracticesProvides a practical security checklist for Claude-assisted development, steering implementations toward least privilege, safe defaults, validated inputs, protected secrets, and review-ready patterns suitable for SaaS, APIs, and agent systems before launch.223installs47Sprint RetrospectiveSprint-retrospective skill guides agents through facilitating sprint retros: gathering what went well, what failed, root themes, and owned action items so teams iterate on process, communication, and delivery habits each cycle.223installs48Testing StrategiesStructured playbook for designing a test strategy: when to use unit versus integration versus end-to-end tests, how to structure fixtures and mocks, and what to enforce in CI before merge. Helps teams avoid either untested releases or slow, brittle suites.223installs49Pattern Detectionpattern-detection in oh-my-skills surfaces repeated bugs, flaky tests, and agent failure modes across sessions, turning noisy traces into grouped patterns with suggested next fixes for faster stabilization.222installs50Agentationagentation helps Claude Code produce and maintain clear documentation for agent systems—roles, tool interfaces, decision flows, and operational notes—so multi-tool agents remain understandable, debuggable, and safe to extend over time.220installs51React GrabHelps developers and agents pull React component structure, props, and JSX from applications to rebuild screens faster, audit UI implementations, and migrate components into new design systems with less manual inspection.220installs52Monitoring ObservabilityGuides Claude through production monitoring and observability for shipped services: instrument code, choose meaningful SLIs, wire metrics and traces, configure actionable alerts, and validate dashboards support fast incident diagnosis.219installs53OhmgMeta game-development skill from oh-my-skills that helps agents plan and coordinate building a shippable game: break epics into tasks, align art, lighting, and store prep work, track dependencies, and keep implementation focused on playable milestones.219installs54Game Performance ProfilerHelps game teams use performance profiling workflows to locate rendering, CPU, and memory bottlenecks in near-release builds, prioritize fixes that improve frame stability, and satisfy platform performance requirements before shipping.218installs55Git SubmoduleGuides agents through git submodule setup, initialization, updates, and common failure modes when projects depend on external repositories. It standardizes vendoring shared packages and keeping nested repos synchronized during active development.218installs56BmadThe bmad skill from akillness/oh-my-skills applies the BMAD (Breakthrough Method for Agile AI-Driven Development) framework so Claude can run structured planning sessions with role-based agent personas. It helps teams validate ideas, define scope, draft PRDs, and decompose work before frontend, backend, or integration build starts.216installs57Deployment AutomationProvides deployment automation guidance for standardizing how builds are promoted, released, and rolled out across environments, reducing manual release errors and accelerating consistent production launches for SaaS, API, and CLI products.216installs58Fabricfabric from oh-my-skills provides pattern-based AI augmentation for agents and CLIs, letting builders chain reusable stencils for research, synthesis, and automation with more consistent, repeatable results.216installs59DebuggingGeneral debugging skill applying a systematic troubleshoot loop: reproduce failures, narrow scope with logs and traces, test hypotheses, ship minimal fixes, and confirm regressions are gone across any codebase layer.215installs60Environment SetupWalks agents through environment setup for projects: choosing runtimes, configuring env files, installing dependencies, containerizing services, and aligning local machines with CI so backends and CLIs run consistently.215installs61Git WorkflowOh-my-skills git-workflow encodes practical Git and PR discipline for shipping: branch models, commit conventions, review-ready diffs, merge/rebase choices, and checklist-driven collaboration for multi-repo development teams.213installs62Firebase Ai LogicFirebase-ai-logic guides integration of Firebase AI Logic with mobile and web clients—secure callables, prompted server rules, auth boundaries, and mapping model outputs into app state and Firestore records.212installs63ScraplingTeaches Scrapling-based scraping workflows: selector design, session handling, pagination, deduplication, and structured exports so teams integrate external web data into agents, APIs, and automation pipelines reliably at build time.212installs64System Environment SetupConfigures system-level development infrastructure—package managers, language runtimes, CLIs, environment variables, and agent prerequisites—so coding agents and local builds run on a consistent, reproducible machine baseline without per-project toolchain drift.212installs65ClawteamClawteam skill from oh-my-skills defines multi-agent team topologies—roles, handoffs, parallel work, and review loops—so Claude Code sessions can run coordinated claw-style agent crews on large tasks.211installs66Firebase CliGuides Claude through Firebase CLI workflows for bootstrapping apps, configuring services, running emulators, and shipping Hosting, Cloud Functions, and database rules safely across environments.211installs67Prompt Repetitionprompt-repetition from oh-my-skills monitors agent dialogs for circular prompts and near-duplicate instructions, then intervenes with de-duplication guidance and recovery steps to keep builds moving efficiently.211installs68Npm Git InstallDocuments and executes correct npm installation from Git repositories—URL formats, ref pinning, semver overrides, and lockfile implications—so agents can wire unpublished forks, monorepo subpaths, or hotfix branches without breaking CI reproducibility.210installs69Skill StandardizationSkill-standardization guides creation and cleanup of Claude Code skills in oh-my-skills, enforcing shared structure, triggers, and documentation so agent instructions remain portable, reviewable, and safe to reuse across repositories and contributors.210installs70LangsmithUses LangSmith to inspect LangChain and LangGraph traces, score outputs, build evaluation datasets, and debug failing tool calls so teams maintain reliable, measurable agent behavior after deployment.209installs71Steam Store Launch OpsCovers Steam store launch operations for games: craft compliant listing copy and media, configure tags and categories, manage depots and branches, plan early access or full release, and execute launch-week updates, patches, and community-facing store changes.209installs72GenkitGuides Claude through building and refining Google Genkit-based AI applications: defining flows, registering tools, configuring models, and aligning with Genkit tracing and evaluation practices for Node and Firebase stacks.208installs73Skill AutoresearchAutomates research to create and refine Claude Code skills: discover relevant sources, extract patterns and gaps, propose skill structures, and iterate drafts so agent capabilities stay current without hand-curating every reference.206installs74HarnessHarness skill standardizes test harnesses for agents and services—fixtures, assertions, regression batches, and CI execution—so teams verify behavior systematically before shipping changes.203installs75Lmstudio CliDocuments LM Studio command-line usage for local LLM development: install tooling, load models, invoke completions, configure endpoints, and embed offline inference into agent prototypes and terminal-driven AI workflows.202installs76Llm Wikillm-wiki from oh-my-skills exposes a structured LLM knowledge base for agents: terminology, model behaviors, prompting tactics, and implementation notes to accelerate informed decisions during discovery and technical research.198installs77Game Build Log TriageTriages game engine build logs from Unity, Unreal, and related toolchains to pinpoint blocking compile, asset, shader, and packaging failures so teams can restore green CI builds before shipping.193installs78Game Demo Feedback Triagegame-demo-feedback-triage structures raw playtest comments, Discord notes, and demo-session reports into actionable buckets for game prototypes. Claude Code can label severity, dedupe themes, and output a ranked backlog so designers and engineers fix blocking issues and clarity problems before the next demo milestone.185installs79Game Ci Cd PipelineSets up a game CI/CD pipeline: configure platform build agents, cache dependencies and assets, run compile plus smoke tests on each PR, package binaries or bundles, publish to staging buckets, and define promotion rules so only green merges reach players or storefront submission.178installs80OooThe ooo skill from oh-my-skills helps solo builders define out-of-office routines for AI coding agents, including what to pause, what to log, and how to hand work back without breaking repos or losing thread.169installs81Improve Codebase ArchitectureImprove Codebase Architecture surfaces architectural friction and proposes deepening refactors that turn shallow modules into deep ones for better testability and AI-navigability. It uses a defined vocabulary (module, interface, depth, seam) and a deletion test to identify pass-through modules, then runs a grilling loop on a chosen candidate. A developer uses it when a codebase is hard to navigate or test and modules feel tightly coupled.164installs82Cavemancaveman from akillness/oh-my-skills is an ultra-compressed communication mode for Claude Code. It strips articles, filler, and pleasantries from agent replies while keeping technical precision, lowering token spend across long implement-debug-review loops.161installs83Compressocompresso from akillness/oh-my-skills condenses bloated agent context—long conversations, tool logs, and pasted documents—into compact summaries that retain critical facts. It helps Claude Code continue complex tasks without hitting context limits or losing thread of prior work.160installs84Stitch Skillsstitch-skills teaches Claude Code how to combine existing skills into coherent pipelines—for example research, then scaffold, then test—so one agent session can reuse vetted SKILL.md modules instead of improvising each step. It is aimed at power users building repeatable agent workflows in oh-my-skills style repos.159installs85Aider Cli WorkflowAider-cli-workflow skill standardizes terminal pair-programming with Aider: starting sessions, context files, bounded edit loops, and commit discipline so agents and developers ship incremental code changes safely across backend, CLI, and agent codebases.154installs86Zoom OutA meta reasoning skill that interrupts tunnel vision: it restates user intent, success criteria, assumptions, and option tradeoffs so agents realign plans, trim scope, and produce clearer next steps instead of drilling deeper into the wrong implementation.154installs87Unity Gamedev Skill Packunity-gamedev-skill-pack from akillness/oh-my-skills bundles Claude guidance for Unity game development, covering scripting, scenes, gameplay systems, and client-side patterns. It targets developers building or extending Unity titles who want consistent C# and editor workflows during active implementation.153installs88Open DesignSupports open, collaborative product design during validation: structure design docs, sketch user flows, define UI direction and interaction patterns, and produce review-ready prototypes that clarify scope before engineering starts.152installs89To Prdto-prd generates a structured Product Requirements Document from existing conversation context and codebase state, synthesizing knowledge rather than interviewing the user. A developer uses it to document requirements for a feature or change and publish it to the project issue tracker before decomposing work into tickets. It matters because it produces a formal, user-facing spec (problem, solution, user stories, out-of-scope) that prevents scope creep.145installs90God Tibo Imagengod-tibo-imagen guides Claude through Imagen-style image generation for product and marketing visuals. It encodes prompt conventions, asset intent, and reuse patterns so agents produce consistent UI imagery, hero art, and campaign graphics during active build work.143installs91PretextHelps authors build effective pretext for AI agents and LLM workflows: frame problems, assemble grounding context, normalize terminology, and scaffold instructions so automated research and coding tasks start from clear, accurate premises.143installs92Ccpi Marketplaceccpi-marketplace from akillness/oh-my-skills connects Claude Code to a CCPI skill marketplace for discovering and installing community skills. It shortens agent setup by letting you browse, compare, and add prebuilt capabilities instead of authoring every SKILL.md yourself.141installs93Github Repo Candidate Quality GateApplies a repeatable quality gate to GitHub repository candidates—checking signals like maintenance cadence, issue hygiene, test presence, and project fit—so teams reject risky deps early and document why a repo was accepted or ruled out.139installs94ClaudekitProvides Claudekit patterns to scaffold Claude Code repos: shared skills, slash commands, hooks, and agent workflows that standardize AI-assisted development setup and reduce bespoke configuration across projects.138installs95ZeudeProvides oh-my-skills agent orchestration patterns for sequencing skills, preserving context, and enforcing guardrails across multi-step coding sessions so autonomous agents complete complex builds more reliably.138installs96GhgrabAutomates fetching structured content from GitHub repositories—such as READMEs, file trees, or targeted paths—so agents can quickly inspect OSS options, examples, and documentation without manual cloning during early research.136installs97Payloadcmspayloadcms teaches Claude how to implement Payload CMS in Node/TypeScript projects: define collections, configure auth and access rules, wire hooks, expose REST/GraphQL APIs, and connect frontends to a typed headless content layer for SaaS and content sites.136installs98Supabase Agent Skillssupabase-agent-skills packages Supabase-specific agent instructions for implementing authentication, relational data access, row-level security, file storage, and serverless endpoints so agents can scaffold and extend Supabase-backed applications correctly.135installs99Hyperfine BenchmarkingTeaches hyperfine CLI benchmarking to measure command latency, compare implementations fairly with warmups, and document perf deltas for scripts, APIs, and build tooling before release.134installs100Openai Agents PythonTeaches Claude how to implement OpenAI Agents in Python, including agent definitions, callable tools, multi-agent handoffs, structured outputs, observability, and safe patterns for shipping reliable autonomous workflows in apps and CLIs.134installs101Grill MeFacilitates a grill-me session where the agent interviews you about a plan or design, probing every branch until requirements, risks, and open questions are explicit enough to commit to implementation confidently.133installs102Setup Pre CommitSets up pre-commit for a repository: install the framework, declare hook repos, map file types to linters/formatters, add secret scanning, and verify hooks block bad commits before push—cutting CI failures and review noise.132installs103Pydantic Aipydantic-ai teaches oh-my-skills users to build production-grade Python agents with Pydantic AI, covering typed tools, validated outputs, composable graphs, and integration patterns for APIs and CLI automations.131installs104Tddtdd encodes test-driven development for Claude: start with failing tests, implement the smallest passing change, then refactor. It steers feature and bugfix work through explicit verification steps so APIs, services, and UI logic stay correct under iterative change.131installs105To Issuesto-issues converts plans, specs, and PRDs into a set of small, independently grabbable tracker issues organized as tracer-bullet vertical slices, helping teams and agents move from ambiguous scope to ordered, actionable backlog work during build planning.131installs106Microsoft Agent Frameworkmicrosoft-agent-framework guides Claude Code through Microsoft Agent Framework setup and implementation: agent structure, tool registration, orchestration patterns, and conventions for reliable LLM-driven assistants in .NET or supported stacks.129installs107Scaffold ExercisesScaffolds coding exercises and practice tasks for agents and developers, enabling quick skill drills, onboarding prototypes, and iterative learning formats before formal curriculum or product work.125installs108Migrate To ShoehornGuides incremental migration of existing backends and agent integrations to the Shoehorn framework, mapping legacy endpoints and configs to Shoehorn modules while preserving behavior through staged refactors and validation gates.124installs109Semblesemble is a fast, token-efficient code search tool for AI agents that returns only the relevant snippets instead of grepping full files. A developer uses it to search a codebase by natural language or symbol, discover semantically similar code, and give agents low-token repo access through an MCP server or CLI. It matters because it cuts search context from roughly 100k tokens to about 2k while keeping high recall.92installs110Browser Harnessbrowser-harness wraps the browser-use/browser-harness project to give an LLM agent a direct Chrome DevTools Protocol connection for multi-step browser workflows like login, navigation, form fill, and data extraction. A developer uses it for autonomous or repeatable browser verification when they want a clean profile instead of an already-open browser. It lets the agent write and repair helpers and includes a Claude-safe screenshot pipeline.70installs111Ultraqaultraqa runs an autonomous QA cycling workflow that tests, diagnoses, fixes, and repeats until the quality goal is met or a concrete blocker is proven. A developer invokes $ultraqa to make tests, build, lint, typecheck, or review gates pass using repo-native commands. It matters because it preserves failure evidence and applies the smallest fix each cycle instead of blindly retrying.55installs112Ultraworkultrawork runs a high-parallelism work burst for independent implementation or cleanup lanes. A developer invokes $ultrawork or $ulw to parallelize separable work such as separate packages, routes, docs, tests, or migrations when shared-file contention is low. It matters because it keeps file ownership explicit and serializes shared edits so parallel lanes integrate cleanly with one final verification pass.54installs113PonytailPonytail pushes the agent to write the least code that fully solves a task by walking a ladder (YAGNI, stdlib, native platform feature, installed dependency, one line, then the minimum that works). It marks every shortcut with a ponytail: comment naming the upgrade path so deferred work becomes a visible ledger. A developer uses it for anti-bloat and YAGNI passes, diff delete-list reviews, or a whole-repo over-engineering audit, with intensity levels lite/full/ultra/off.53installs114AgenticskillsAgenticSkills is a thin wrapper around the official oh-my-gods install.sh that copies the whole bundle of 80+ agent skills onto a machine in one step. Use it when a new machine needs the full catalog mirrored into Claude Code, Codex CLI, Gemini, and OpenCode, or to refresh an existing install. It pins the installer URL and exposes environment knobs like PLATFORM and WITH_LANGCHAIN.50installs115AutopilotAutopilot is an exact-name entry point for a Codex/OMX-style autonomous build workflow that takes a product or code request through planning, implementation, QA, and final verification. A developer invokes it with $autopilot when they want the agent to build the whole thing end to end rather than a focused one-file edit. It implements in small safe slices and stops on repeated failures or missing decisions.43installs116Spec Kitspec-kit is a routing-first wrapper around GitHub's specify CLI for Spec-Driven Development. A developer uses it to install specify-cli, initialize a project for a chosen AI coding agent, and drive the constitution, specify, plan, tasks, and implement command pipeline so specs become executable and drive implementation. It matters because it lets a single executable spec be shared across 30+ agents before code is written.30installs117Cli Anythingcli-anything is a routing wrapper for the HKUDS CLI-Anything ecosystem that lets AI agents drive real desktop and server software through structured Click CLIs rather than brittle GUI automation. A developer uses it to install ready-made harnesses via CLI-Hub, give agents an autonomous discovery meta-skill, or generate a new harness from a codebase through a 7-phase pipeline. It targets apps like GIMP, Blender, LibreOffice, and Godot.26installs118Opikopik wraps Comet's open-source Opik to give a developer LLM observability, evaluation, and optimization from one skill. It installs the SDK, stands up a server on cloud, Docker Compose, or Kubernetes, wires tracing through @opik.track or 50+ framework integrations, and scores outputs with LLM-as-a-judge metrics. Use it when you need LLM tracing, prompt evaluation, production monitoring, or guardrails.26installs119Spec Stackspec-stack composes GitHub spec-kit, Ouroboros (ooo), and HKUDS CLI-Anything into a single spec-driven delivery stack that goes from a written spec to a verified artifact. A developer uses it when spec authoring, an immutable verified loop, and real software harnesses all need to work together without fighting over the same job. It matters because it hands a concrete artifact between layers (spec feeds the seed, the seed names the harnesses, harness JSON output is evaluate evidence).17installs120Deep DiveDeep Dive is a two-stage pipeline that first investigates why something happened by tracing causal hypotheses, then defines what to do about it through a deep-interview requirements process. It runs three parallel causal investigation lanes and injects the findings into the interview so context is not lost between steps. A developer uses it when a problem is ambiguous and needs investigation before requirements, producing a runtime-portable spec.1installs121DeepinitDeep Init creates or refreshes hierarchical AGENTS.md documentation across a codebase so agents can understand what each directory contains and how components relate. It generates parent-referenced AGENTS.md files level by level, preserves manually added notes on regeneration, and excludes runtime state directories like .omc and .codex. A developer uses it to make a repo AI-navigable and keep its documentation current.1installs122Drawiodrawio wraps the Agents365-ai drawio-skill to convert natural-language descriptions into editable .drawio diagrams and export them to PNG, SVG, PDF, or JPG via the draw.io desktop CLI. It can also turn an existing codebase into an auto-laid-out structure diagram. A developer uses it for precise architecture, ERD, UML, sequence, or flowchart diagrams with real vendor icons, driven through a plan, generate, self-check, iterate, export loop.1installs123React Best Practicesreact-best-practices is a Vercel-authored skill for measurement-led React and Next.js performance audits. A developer uses it to diagnose or refactor a slow React UI, App Router route, or heavy client component by first classifying the dominant bottleneck (waterfalls, bundle size, server/client boundary, rerender churn, hydration cost) and then applying prioritized fixes. It matters because it routes work to the right rule packet instead of guessing at optimizations.1installs124Teamteam is an exact-name skill that runs or routes a coordinated multi-agent team workflow for Codex/OMX when the user invokes $team. A developer uses it when workers need shared runtime state or a shared task list, backed by tmux panes and omx. It matters because it enforces real runtime evidence (team name, panes, state path) instead of simulating fan-out.1installs125Typesensetypesense is a routing-first skill for standing up a self-hostable, typo-tolerant search backend with Typesense, an open-source Algolia and ElasticSearch alternative. A developer uses it to pick a server mode (binary, Docker, or Typesense Cloud), install an API client, design a collection schema, index documents, and run searches with faceting, geo-search, synonyms, and scoped API keys. It matters when adding site, app, or product search or migrating off Algolia/Elasticsearch.1installs126Vercel Deployvercel-deploy runs Vercel-specific deployment operations for linked web and fullstack projects: preview deploys, production deploys, staged deploy-and-promote flows, aliases and domains, environment-variable sync, and rollback response. A developer uses it when the provider is already Vercel and they need a preview URL, stable alias, custom domain, env-scope fix, production cutover, or rollback. It matters because it picks one deploy mode and verifies with the right depth rather than dumping every command.1installs127Web Design Guidelinesweb-design-guidelines runs a broad web UI review for pages, flows, or component surfaces across audit categories such as layout hierarchy, content clarity, CTA emphasis, visual consistency, interaction states, responsiveness, and accessibility basics. A developer uses it to turn a vague design-feedback request into a structured, prioritized audit before launch. It matters because it separates broad interface issues from specialist remediation lanes and produces one concise audit packet ordered by friction and launch risk.1installs