🎯Skills75
Enables Google Image search capabilities within Claude Code, from a collection of skills extending AI with specialized workflows and tools.
Skill for generating PDF documents, part of a Claude skills collection that also provides Apple Health data querying and analysis capabilities.
FireCrawl research skill for enriching documents with web sources by automatically searching and scraping content, generating BibTeX bibliographies, and providing academic writing templates
Skill for querying and analyzing Apple Health data from a local SQLite database covering vitals, activity, sleep, and workouts with support for Markdown, JSON, and FHIR R4 output formats.
HTML presentation generator skill with neobrutalism styling, ASCII art decorations, smooth scrolling navigation, keyboard controls, and export to PNG and PDF formats
Generate structured decision-making tools including step-by-step guides, bias checkers, scenario explorers, and interactive dashboards.
ElevenLabs text-to-speech skill for generating professional audio from text with customizable voice parameters, pre-configured voice presets, and direct script execution
Deep research skill that conducts comprehensive internet-enabled research using OpenAI Deep Research API, with automated prompt enhancement through clarifying questions and web search capabilities.
Telegram messaging skill for fetching, searching, downloading, sending, and editing messages on Telegram from Claude Code.
Applies Agency brand colors, typography, and neobrutalism design style to artifacts including presentations, SVG graphics, documents, and web interfaces for consistent brand identity.
A Claude Code skill for evidence-based health research using tiered trusted sources with GRADE-inspired evidence ratings, integrating Apple Health data for personalized context on health, nutrition, exercise, and wellness questions.
Analyzes meeting transcripts using Cerebras AI (llama-3.3-70b) to extract decisions, action items, opinions, questions, and terminology.
A skill for searching, fetching, and downloading emails from Gmail, supporting queries like searching for emails from specific senders, filtering unread messages, and downloading attachments.
Fetches meeting recordings, transcripts, summaries, and action items from Fathom API.
A test-driven development skill that enforces disciplined RED-GREEN-REFACTOR cycles using multi-agent orchestration, where a Test Writer agent never sees implementation code and an Implementer agent never sees the specification. Supports Jest, Vitest, pytest, Go test, cargo test, PHPUnit, and RSpec.
Publishes files or Obsidian notes as GitHub Gists, enabling quick creation of shareable code snippets and markdown content directly from your editor.
A skill for creating and managing Zoom meetings and accessing cloud recordings via the Zoom API, supporting meeting creation, listing, and scheduling.
A skill that extracts YouTube video transcripts with metadata using yt-dlp and saves them as formatted Markdown with YAML frontmatter to an Obsidian vault for personal knowledge management.
Skill for processing textual and multimedia files with various LLM providers using the llm CLI, supporting both non-interactive and interactive modes with model selection and config persistence.
A skill for querying Chrome browsing history with natural language, enabling filtering by date range, article type, keywords, and specific sites.
A command handler skill for executing Claude Code commands in a Telegram agent project, used for fixing bugs, adding features, or modifying bot code.
A meeting importer skill that queries Granola's Personal API to list, view, and export meeting notes and transcripts to Obsidian markdown in a Fathom-compatible format, including AI-generated summaries and speaker-attributed transcripts.
A Claude Code skill enabling constructive, evidence-based dialogue with multiple analysis modes (full, interactive, TL;DR, steelman, decision), designed to avoid sycophancy and provide genuine critical feedback.
Generates interactive AI transformation context-building prompts for consulting clients, guiding structured discovery sessions about business pain points, tech stack, and AI opportunities.
Analyzes Wispr Flow voice dictation history from the local SQLite database for self-reflection, work patterns, mental health insights, and productivity analytics across bilingual dictations.
CLI skill for interacting with Google Workspace services including Gmail triage and send, Calendar agenda and events, Drive upload and share, Tasks, Docs, and People/Contacts via the gws binary.
Generates and edits images using Google's Gemini image generation models (Nano Banana family), supporting style presets, platform-specific sizing, variants, reference images for style transfer, and organized output with metadata.
Skill for generating 3D interactive knowledge maps from Obsidian vaults as single HTML files using Three.js, with concentric entity rings, multi-scale semantic zoom, and dual-graph comparison between vaults.
Interactive image generation and editing via OpenAI's GPT Image 2 API with style presets, platform-aware sizing, and a cost-saving draft-to-final workflow. Supports single images, photo edits, LinkedIn/Instagram carousels, and variant generation.
Processes meeting transcripts by auto-detecting the meeting type (leadgen, partnership, coaching, internal) and extracting type-specific structured analysis including commitments, action items, deal assessments, and sentiment via LLM-powered extraction.
Searches Claude Code session transcripts using keyword pre-filtering and semantic evaluation, supporting natural language queries with synonym matching to find previous conversations about specific topics.
A collection of 20+ Claude Code skills spanning health research, deep research automation, Telegram/Gmail integration, browsing history search across synced devices, YouTube transcript extraction, PDF generation, presentation creation, and more.
A demo/recording mode skill that redacts personally identifiable and sensitive information from Claude Code outputs in real time, replacing names, emails, locations, dates, financial values, medical info, and credentials with obviously fake placeholder data during screen-shares and demos.
A final retrospective and self-assessment skill for Claude Code Lab graduates, running four sequential interactive parts: progress audit, best prompt capture, monthly plan creation, and structured feedback collection.
A terminal-first Jobs-to-Be-Done engine that helps founders and product teams conduct structured interviews, capture switching forces (Push/Pull/Habit/Anxiety), score opportunities with ODI, and export decision-grade artifacts including JSON briefs and messaging angles.
Creates standalone HTML data reports and infographic dashboards inspired by Edward Tufte's design principles, featuring interactive Chart.js visualizations, inline SVG sparklines, EB Garamond typography, and responsive 2-column narrative+data layouts.
Interview-driven automation design tool that conducts structured JTBD (Jobs-to-be-Done) interviews via text or voice, then exports a one-page markdown spec and SVG design map. Supports multiple interview styles, coverage tracking, and session seeding from prior transcripts.
Scores and compares images using vision LLMs as judges with YAML-defined criteria presets for 11 use cases including text-to-image, photorealism, document OCR, and UI evaluation.
Standalone CLI for the Linear issue tracker with browser OAuth authentication, enabling creation, listing, updating, and searching of Linear issues and project management.
Indexes and searches Claude Code sessions using Gemini semantic embeddings, enabling users to find past sessions by topic and relaunch the best match with cosine similarity ranking.
Controls the timeBuzzer hardware LED over USB MIDI, supporting color commands, effects (pulse, strobe, rainbow, fade), semantic status signals, and per-segment RGB control across its 3 LED segments. Can be synced with Hue lighting using a shared signal vocabulary.
A skill for three-layer PII anonymization of session transcripts using Natasha (Russian NER), OpenAI Privacy Filter, and local LLM (Ollama) in sequence, designed for therapy, coaching, and consulting records with fully local processing.
Automates getSorted.de for German freelancer tax and invoicing workflows — creates and sends invoices, tracks expenses, and submits VAT, EU summary, and annual tax returns to the Finanzamt via ELSTER, using real-browser automation with Chrome Beta.
Runs OpenAI Codex CLI commands for code analysis, refactoring, and automated editing within a sandboxed environment configured with a ChatGPT account.
Iterates on RAG systems with structured evaluations instead of eyeballing, providing ranked experiments with cost tracking and feedback on retrieval strategies.
A Telegram communications assistant skill that reviews pending response drafts, manages an approval queue for outgoing messages, and tracks follow-ups for unanswered outbound messages with exponential backoff reminders.
A Claude Code skills collection including evidence-based CBT and DBT cognitive toolkit with guided thought records, opposite action, DEAR MAN roleplay, and crisis intervention.
Generates a daily digest of 3-7 genuinely surprising items from newsletters and Telegram channels by scoring content for epistemic friction against the user's current focus and research, then appending selected items to a daily note.
A session retrospective skill for continual learning that reviews conversations, extracts successes, failures, and discoveries, then updates skill files with dated learnings to compound knowledge over time.
A pre-consultation diagnostic questionnaire skill for clients building websites with AI tools (Claude, Cursor, Bolt, v0) who have concerns about quality, design, or maintainability. It walks through structured interview rounds and generates a consultation brief in Russian or English.
A Claude Code skill that audits your Claude Code ecosystem health, checking skill staleness, project activity, CLAUDE.md instruction drift, and Mac Mini service status.
Generates self-contained HTML presentations with dual article/slides mode, ElevenLabs voiceover narration synced to slides, scroll-reveal animations, and optional GPT Image 2 illustrations.
An end-to-end automation pipeline for publishing Claude Code lab meeting recordings, handling transcript extraction, video download from Zoom/Fathom, YouTube upload, documentation site generation, and thumbnail creation in a single sequential workflow. Supports configurable paths and automatic language detection for Russian and English transcripts.
Generates a fully structured Obsidian learning vault for any certification, course, or study goal. Creates organized notes with domains, concepts, lessons, scenarios, Maps of Content, Dataview queries, action items, and multiple navigation paths.
A research-to-publication skill that produces scored, source-verified weekly industry digests by casting a wide net via web search, verifying sources for AI-generated slop, scoring items on novelty, relevance, and technical depth, then generating markdown and Tufte-style HTML reports.
A Claude Code skill that automates the end-to-end pipeline for publishing meetup recordings to YouTube, including downloading from Zoom, adding intro/outro sequences, generating thumbnails, creating descriptions with timecodes, and uploading.
A skill for semantically searching local Obsidian vaults and markdown knowledge bases using the on-device qmd engine, offering BM25 keyword, vector similarity, hybrid, grep, and fused search modes with cross-language retrieval.
Automates the full pipeline for publishing skills to a claude-skills-site Astro website, including reading SKILL.md, generating MDX frontmatter and body, selecting the right bundle, creating hero images, and committing changes to both repos.
Generates social media covers and visual assets for AGENCY Community events, meetups, and YouTube recordings using an established design system with specific typography, colors, and layout rules.
Prepares comprehensive meeting prep notes by pulling Cal.com bookings, researching participants via web search, and auditing previous session history from an Obsidian vault. Automatically links prep notes to post-meeting session notes.
Generates realistic, persona-consistent synthetic coaching and therapy session transcripts for evals, demos, and training data. Supports multiple modalities (ICF/GROW, CBT, IFS, ACT) and output formats including Fathom/Granola style, plain dialogue, structured JSON, and Obsidian markdown.
Analyzes coaching or therapy session transcripts and appends structured summaries with key insights, decisions, action items, and cross-session pattern detection. Runs entirely within Claude Code with no external API calls.
A skill for driving the Cull app, a local AI-art image library for importing, browsing, rating, organizing, and exporting images with vision/quality analysis and embedding-based similarity search, controllable via CLI, URL scheme, GUI, or MCP server.
Part of a Claude Code skills collection by Gleb Kalinin that includes specialized workflows for software release management with tiered compatibility policies and synthetic coaching session transcript generation with multiple modalities and persona consistency.
A skill for coordinating Claude Code, Codex, and similar AI agents as peer experts rather than in a hierarchy, defining when to hand off tasks between agents based on their operating surfaces, tools, and evidence access.
A collection of Claude Code skills offering specialized workflows including tiered release management with compatibility policies, synthetic session generation, name audition for brand research, and peer agent collaboration.
A skill for auditioning candidate product, brand, or company names by running domain availability checks, collision research across SaaS/GitHub/packages, trademark reads, and producing a ranked casting report of finalists.
A skill that converts Python CLI scripts into agent-friendly NDJSON output by adding a --json flag, or scaffolds a complete cli_utils package for machine-readable structured output.
Skill
Skill
Skill
Skill
Skill
Skill
Skill