daymade/claude-code-skills
64 skills46.1k installs84.2k starsGitHub
Install
npx skills add https://github.com/daymade/claude-code-skillsSkills in this repo
1Twitter ReaderFetch Twitter X post and article content with full media support For X Articles with images use the new fetch_article py script bash uv run with pyyaml python scripts fetch_article py article_url output_dir Example bash uv run with pyyaml python scripts fetch_article py https x com HiTw93 status 2040047268221608281 Clippings This will Fetch structured data via twitter cli likes retweets bookmarks Fetch content with images via jina ai API Download all images to attachments YYYY MM DD AUTHOR TITLE Generate complete Markdown with embedded image references Include YAML frontmatter with metadata The twitter reader agent skill provides documented workflows prerequisites triggers and safety guidance from its SKILL md source Agents load it when user requests match the description and follow step by step instructions without inventing capabilities It integrates with standard agent tooling for the tasks inputs outputs and failure modes described in the repository documentation2.4kinstalls2I18n ExpertThe i18n-expert skill is designed for this skill should be used when setting up, auditing, or enforcing internationalization/localization in UI codebases (React/TS, i18next or similar, JSON locales), including. I18n Expert Overview Deliver a complete i18n setup + audit pass: configure the i18n framework, replace user-facing strings with keys, ensure locale parity, and validate pluralization/formatting for en-US and zh-CN. Core Capabilities Library selection and setup (React, Next.js, Vue). Invoke when the user asks about i18n expert or related SKILL.md workflows.1.7kinstalls3Prompt OptimizerThe prompt-optimizer skill converts vague prompts into precise specifications using EARS Easy Approach to Requirements Syntax from Rolls-Royce methodology. Its four-layer process applies EARS syntax transformation, domain theory grounding with frameworks like GTD or BJ Fogg, example extraction with real data, and structured Role Skills Workflows Examples Formats output. A six-step workflow analyzes weaknesses such as broad scope, missing triggers, ambiguous actions, and absent constraints, then rewrites requirements into testable when-then patterns. It triggers on optimize my prompt, improve this requirement, or loose feature descriptions lacking measurable outcomes. Use when developers need requirement refinement before AI-generated code, products, or documents.1.5kinstalls4Ppt CreatorThe ppt-creator skill is designed for create professional slide decks from topics or documents. Generates structured content with data-driven charts, speaker notes, and complete PPTX files. Applies persuasive. PPT Creator > Goal: Transform a simple topic into a presentation-ready, high-quality slide deck. When key information is missing, use the minimal intake form (references/INTAKE.md) to gather context or apply safe defaults. Invoke when the user asks about ppt creator or related SKILL.md workflows.1.4kinstalls5Qa ExpertThe qa-expert skill this skill should be used when establishing comprehensive QA testing processes for any software project Use when creating test strategies writing test cases following Google Testing Standards executing test plans tracking bugs with P0-P4 classification calculating quality metrics or generating progress reports Includes autonomous execution capability via master prompts and complete documentation templates for third-party QA team handoffs Implements OWASP security testing and achieves 90 QA Expert Establish world-class QA testing processes for any software project using proven methodologies from Google Testing Standards and OWASP security best practices When to Use This Skill Trigger this skill when Setting up QA infrastructure for a new or existing project Writing standardized test cases AAA pattern compliance Executing comprehensive test plans with progress tracking Implementing security testing OWASP Top 10 Filing bugs with proper severity classification P0-P4 Generating QA reports daily summaries weekly progress Calculating quality metrics pass rate coverage gates Preparing QA documentation for third-party team handoffs Enabling autonomous LLM-driven test ex.1.3kinstalls6Macos Cleanermacos-cleaner in daymade/claude-code-skills guides developers through reclaiming tens of gigabytes on a development Mac without breaking installed applications. The skill documents cleanup targets with explicit safety levels—~/Library/Caches for browser and app caches is marked safe to delete, with notes that first launches may be slower after cache removal. A security scan passed using gitleaks plus pattern-based validation (scanned 2026-06-05). Developers reach for macos-cleaner when Xcode, Docker, npm, or browser caches have consumed large chunks of disk and manual Finder cleanup is risky or incomplete. The reference explains what each target stores, expected impact after deletion, and which paths are safe versus cautionary. Triggers include low disk space warnings, bloated ~/Library/Caches, and routine Mac maintenance before installing large SDKs or containers.1.2kinstalls7Deep Researchdeep-research is a Claude Code skill from daymade/claude-code-skills that orchestrates thorough, multi-source investigations producing structured evidence rather than shallow AI summaries. The skill includes a completeness review checklist covering required sections, heading format, length targets, citation coverage, Tier A/B sources for numeric claims, conflicting-source notes, and evidence tables mapping claims to sources. A bundled security scan note records gitleaks and pattern-based validation with a content hash dated 2026-01-25. Developers reach for deep-research when due diligence, competitive analysis, or technical fact-finding must be defensible with citations and explicit conflict handling before recommendations ship.1.1kinstalls8Ima Copilotima-copilot is a one-command installer and repair skill for the upstream Tencent ima-skill wrapper across Claude Code, Codex, and OpenClaw. It orchestrates ima-skill setup, diagnoses invalid SKILL.md frontmatter errors in submodule skill files, and configures IMA API credentials, fan-out search, preferred knowledge-base priority boosting, and 笔记搜索 note search. Developers reach for ima-copilot when mentions of IMA, 腾讯 IMA, ima.qq.com, 知识库搜索, or ima-skill install failures appear during agent configuration. The skill explicitly states it is a wrapper layer that orchestrates upstream ima-skill rather than replacing Tencent's official skill. Typical triggers include Skipped loading skill(s) due to invalid SKILL.md startup errors and missing YAML frontmatter bugs in nested ima-skill repositories. ima-copilot personalizes search routing so multi-agent coding workflows can query enterprise knowledge bases reliably without manual submodule debugging, credential guesswork, or repeated install retries.1kinstalls9Ui Designerui-designer is a daymade claude-code-skills agent skill that acts as a collaborative SaaS product consultant, turning application ideas marked as IDEA or MVP into comprehensive project specifications. Each user reply gets integrated into a repeating structured plan covering elevator pitch, problem statement, target audience, unique selling proposition, target platforms, categorized feature lists with user-story checkboxes, and UX/UI state considerations. Despite the name, the skill focuses on product specification and MVP scoping rather than visual UI implementation. Developers reach for ui-designer when they have a rough app concept and need a detailed, iteratively refined spec before writing code.1kinstalls10Fact Checkerfact-checker verifies factual claims in documents using web search and authoritative sources, then proposes corrections with user confirmation before any edits apply. It triggers on requests to fact-check, verify information, validate claims, check accuracy, or update outdated data in documentation. Supported claim types include AI model specifications, technical documentation, statistics, and general factual statements. Developers reach for fact-checker before publishing README files, blog posts, or internal docs where stale model specs or incorrect version numbers would mislead readers.980installs11Developing Ios Appsdeveloping-ios-apps is a Claude Code skill from daymade/claude-code-skills for end-to-end Apple platform development. It covers XcodeGen project.yml generation, SwiftUI interfaces, Swift Package Manager dependency resolution, Apple Developer code signing, notarization with notarytool, and GitHub Actions CI/CD including secrets in conditionals. The skill triggers on concrete failure modes: Error -25294, keychain mismatch, adhoc fallback, EMFILE, notarization credential conflicts, continueOnError, Library not loaded @rpath, iOS version compatibility, and Electron @electron/osx-sign/@electron/notarize configuration. Developers reach for it when building for real devices, fixing Xcode build failures, or wiring certificate and provisioning workflows that block release.968installs12Promptfoo Evaluationpromptfoo-evaluation is a Claude Code skill that maps Promptfoo provider configuration for systematic LLM output testing across prompts, models, and configs. The reference covers echo providers for zero-cost prompt previews, Anthropic messages providers such as claude-sonnet-4-6 with max_tokens and temperature, and patterns for comparing runs before committing prompts. Developers use it when validating few-shot structure, debugging variable substitution, and scoring candidate prompts without burning tokens on exploratory runs. The bundled API reference emphasizes YAML-first setup and security-validated content suitable for agent-driven evaluation loops inside Claude Code.968installs13Meeting Minutes Takermeeting-minutes-taker is a Claude Code skill that systematically reviews AI-generated meeting minutes against source transcripts using a four-phase completeness checklist. Phase 1 checks structure—topics, attributed decisions, quoted support, and named action-item owners with priorities. Phase 2 verifies content depth per transcript section. The skill documents common omission patterns for technical, business-logic, and process items, plus merge rules for multi-version minutes and red flags for missing content. Developers reach for meeting-minutes-taker after standups, architecture reviews, or incident calls when draft minutes may drop speakers, decisions, or concrete next steps.966installs14Skill Creatorskill-creator implements a Post-hoc Analyzer Agent in daymade/claude-code-skills that runs after blind comparator tests determine a winner. Given winner label A or B, winner skill path, winner transcript path, and loser counterparts, the analyzer unblinds results by examining skills and execution transcripts. The goal is actionable insight: what made the winner better and how to improve the losing skill. Developers reach for skill-creator when iterating on agent SKILL.md quality after A/B blind tests rather than during initial skill authoring.961installs15Mermaid Toolsmermaid-tools is a skill from daymade/claude-code-skills that automates Mermaid diagram rendering in codebases and documentation. It requires Node.js and a global @mermaid-js/mermaid-cli install verified with mmdc --version, then uses bundled extract-and-generate scripts to pull Mermaid fenced blocks from files and export validated PNG or SVG assets. Developers reach for mermaid-tools when README or docs contain ```mermaid blocks that must become checked-in images for PDFs, wikis, or PR review. The skill documents environment variables, WSL2 Chrome setup for headless rendering, and a validation workflow to catch broken diagram syntax before publish.932installs16Skill Reviewerskill-reviewer is a Claude Code skill that evaluates custom SKILL.md files against a published best-practices checklist before sharing or publishing. It checks YAML frontmatter rules—name max 64 characters, lowercase hyphens only, description max 1024 characters in third-person voice with trigger conditions—and reviews description quality, imperative step wording, and progressive disclosure patterns. A bundled security scan (gitleaks plus pattern validation, scanned 2026-01-15) confirms the skill content passes secret-detection gates. Developers reach for skill-reviewer when a new or edited skill needs a structured preflight audit instead of informal peer review.927installs17Excel Automationexcel-automation is a macOS-focused skill from daymade/claude-code-skills for creating, parsing, and controlling Excel files programmatically. Developers run uv scripts such as create_formatted_excel.py with openpyxl for professional report formatting, and parse_complex scripts using stdlib zipfile and XML when openpyxl cannot handle investment-bank xlsm financial models. AppleScript commands automate Excel window control on macOS. Reach for excel-automation when generating formatted .xlsx reports, extracting data from complex .xlsm models, or driving Excel UI automation on a Mac rather than manual spreadsheet work.919installs18Statusline Generatorstatusline-generator is a Claude Code statusline skill that surfaces real-time API spend directly in the terminal prompt by calling the ccusage command-line tool. The statusline script invokes `ccusage session --json --offline` and `ccusage daily --json --offline`, parsing totals with jq to show session and daily cost figures formatted to two decimal places. Developers reach for statusline-generator when running long Claude Code sessions and need immediate cost feedback—avoiding post-hoc transcript audits or surprise daily bills. The skill includes a ccusage integration reference for troubleshooting offline JSON parsing and prompt formatting.916installs19Github Opsgithub-ops is a skill from daymade/claude-code-skills that documents GitHub REST and GraphQL operations accessible through the gh api CLI. It covers authentication via gh auth login, pull requests, issues, repositories, Actions and workflows, search, GraphQL queries, rate limiting, and webhooks with copy-ready command patterns. Developers reach for github-ops when agents must query or mutate GitHub state—listing PRs, triaging issues, inspecting workflow runs, or searching code—without leaving the terminal session. The skill assumes an authenticated gh token and focuses on practical API surfaces engineers hit during repo maintenance and automation.911installs20Youtube Downloaderyoutube-downloader is an internal SOP skill from daymade/claude-code-skills that standardizes yt-dlp downloads inside AI agent shell workflows. The playbook enforces quoted YouTube URLs in zsh where ? triggers glob expansion, aligns HTTP_PROXY and HTTPS_PROXY across yt-dlp and PO Token providers, handles bot-check prompts with browser cookies, and starts PO Token providers—preferring Docker bgutil with browser fallback. A gitleaks security scan is documented in the skill source. Developers reach for youtube-downloader when automated video or audio extraction fails mid-agent-run and manual trial-and-error would waste turns on known failure modes.907installs21Skills Searchskills-search is a Claude Code skill from daymade/claude-code-skills that auto-bootstraps agents to search, discover, install, and manage skills from the CCPM registry. Triggers include requests to find skills for PDF, code review, or cloud troubleshooting, install named skills like cloudflare-troubleshooting, or list what is already installed locally. The skill uses Bash and Read tools with behavioral directives so agents query the registry and perform installs in-session. Developers reach for skills-search when expanding agent capabilities mid-task instead of manually cloning skill repositories or guessing available packages.896installs22Claude Code History Files Finderclaude-code-history-files-finder is a claude-code-skills marketplace skill with a 314-line SKILL.md, two Python scripts, and JSONL format references for Claude Code session files. Scripts analyze_sessions.py searches and analyzes session history while recover_content.py extracts recoverable content from stored JSONL records documented in references/session_file_format.md. Developers reach for this skill when prior Claude Code work disappears from the editor but may still exist on disk in hidden session stores. The skill passed security validation with a .security-scan-passed marker and follows marketplace conventions for executable recovery tooling.895installs23Capture Screencapture-screen is a macOS programmatic screenshot skill with a three-step workflow: find windows, control views, capture images. It resolves window IDs using a Swift script built on CGWindowListCopyWindowInfo, controls application windows through AppleScript for zoom, scroll, and selection, then captures PNGs with screencapture using the -x -l window-id flags. Example commands include swift scripts/get_window_id.swift Excel followed by screencapture -x -l 12345 output.png. Developers reach for capture-screen when automating documentation screenshots, building multi-shot visual workflows, or running visual QA on macOS apps. The skill requires macOS tooling and does not apply to Linux or Windows capture tasks.775installs24Docs Cleanerdocs-cleaner is an Agent Skill from daymade/claude-code-skills that applies a value analysis template to project documentation section by section. Each section is scored for Keep, Delete, or Condense with line counts and reasons—for example keeping essential setup instructions while deleting content that duplicates CLAUDE.md. Developers reach for docs-cleaner when README sprawl, overlapping guides, or verbose technical sections make onboarding slow. The template tracks unique troubleshooting steps, frequently referenced procedures, and technical debt notes that must survive consolidation.757installs25Claude Md Progressive Disclosurerclaude-md-progressive-disclosurer is a CLAUDE.md optimization skill grounded in progressive disclosure principles for coding agents. The goal is maximizing information efficiency, readability, and maintainability by keeping high-signal rules in CLAUDE.md while moving depth and citations into references/. The skill itself models the pattern: core methodology stays in SKILL.md while detailed examples sink to references, targeting SKILL.md at 500 lines per Anthropic skill guidance. Developers invoke it when CLAUDE.md grows duplicated, agents repeatedly ignore rules, or information is scattered across files. The skill forbids using line count alone as a success metric, emphasizing single source of truth, cognitive relevance, and trigger-based rule placement instead of aggressive deletion that removes necessary guardrails.755installs26Pdf Creatorpdf-creator is a Claude Code skill from daymade/claude-code-skills that generates clean, professional PDF documents from markdown or structured content using Claude. Listed on skills.sh with 536 installs, it targets developers who need polished exports without building a custom PDF pipeline. The skill guides the agent through formatting and output so READMEs, specs, and reports become presentation-ready PDFs. Reach for pdf-creator when documentation already exists as markdown and you need a distributable PDF artifact quickly.733installs27Transcript Fixertranscript-fixer is a Claude Code skill from daymade/claude-code-skills listed on skills.sh with 544 installs and catalog rank 8964. The skill repairs meeting and recording transcripts by correcting misheard words, fixing speaker attribution, removing filler, and restoring readable sentence structure for downstream documentation. Developers reach for transcript-fixer after generating raw ASR output from standups, interviews, or recorded demos when the text is too noisy for specs, tickets, or meeting notes. The skill focuses on editorial cleanup rather than summarization or action-item extraction.732installs28Github Contributorgithub-contributor is a phase-based Claude Code skill from daymade/claude-code-skills for contributing upstream to third-party GitHub repositories. The playbook spans repository discovery, CONTRIBUTING.md compliance, PR size checks, minimal-diff implementation, PR descriptions with AI-assisted disclosure, rebase and conflict resolution, and post-submission maintainer interaction. Developers reach for github-contributor when opening or fixing an upstream PR, rebasing against main, or responding to maintainer or bot review feedback on an owner/repo they do not control. Triggers include phrases like submit a PR, fix this upstream, rebase against main, and respond to the bot review across English and Chinese contexts.717installs29Llm Icon Finderllm-icon-finder is a Claude Code skill from daymade/claude-code-skills that locates official AI and LLM brand icons via the lobe-icons library on GitHub. The library contains 100+ icons for models like Claude, GPT, and Gemini, providers like OpenAI, Anthropic, and Google, and applications like ComfyUI and LobeChat. Developers invoke it when building model selector UIs, comparison tables, or docs pages that need correct brand assets in SVG, PNG, or WEBP with light and dark theme variants. The skill returns icon URLs and format guidance so agents skip manual logo hunting and trademark-inconsistent placeholders.715installs30Video Comparervideo-comparer is a daymade/claude-code-skills utility that compares two video files—typically an original and a compressed export—and produces an interactive HTML report analyzing compression quality. The bundled script extracts video metadata, calculates objective metrics including PSNR and SSIM, and renders frame-by-frame visual comparisons with multiple viewing modes such as slide-based inspection. Triggers include compare videos, video quality, compression analysis, before/after compression, and requests to assess encoded output fidelity. Developers reach for video-comparer when tuning bitrate, codec, or transcoding presets and need quantified evidence plus side-by-side frames instead of subjective playback checks alone. The skill fits content pipelines, media CLI tooling, and pre-release validation where degraded compression must be caught before publishing or distribution.712installs31Teams Channel Post Writerteams-channel-post-writer is a daymade/claude-code-skills agent skill that drafts educational Microsoft Teams channel posts teaching colleagues how to use Claude Code features, tools, and best practices. The skill provides templates, writing guidelines, and structured approaches emphasizing concrete examples, underlying principles, and context engineering practices. Developers reach for teams-channel-post-writer when announcing new Claude Code capabilities, sharing productivity tips, documenting lessons learned, or publishing internal feature guides inside Teams channels. Outputs are formatted for internal knowledge sharing rather than public SEO articles, making them suitable for engineering enablement and rollout communications.710installs32Cli Demo Generatorcli-demo-generator is an agent skill from daymade/claude-code-skills that records professional animated CLI demos as GIFs with VHS terminal tapes. The workflow handles tape file creation, self-bootstrapping demos with hidden setup steps, output noise filtering, post-processing speed-up, and frame-level verification. Four approaches span fully automated to manual tape editing. Developers reach for cli-demo-generator when README files need animated proof of a CLI workflow, product pages must showcase shell commands, or requests mention "record terminal", "VHS tape", "demo GIF", or "animate my CLI". The skill targets documentation and marketing visuals for command-line tools rather than changing tool behavior.707installs33Cloudflare Troubleshootingcloudflare-troubleshooting is an evidence-driven Cloudflare debugging skill built on falsification, layered isolation, and counter-review methodology from the debugging-network-issues base skill. It instructs agents to query the Cloudflare API for actual configuration state instead of guessing when users hit ERR_TOO_MANY_REDIRECTS, SSL certificate errors, DNS mismatches, or other edge issues. The skill layers Cloudflare-specific investigation on top of general network diagnosis discipline. Developers reach for cloudflare-troubleshooting when production traffic fails at the Cloudflare edge and dashboard assumptions need API-backed verification.707installs34Claude Skills Troubleshootingclaude-skills-troubleshooting is a daymade/claude-code-skills agent skill with 524 installs listed on skills.sh that guides systematic diagnosis when Claude Code, Cursor, or custom agent skills fail to load, trigger, or execute correctly. Developers reach for claude-skills-troubleshooting after skill manifests, paths, hooks, or agent configurations break following upgrades or repo changes. The skill focuses on reproducible troubleshooting steps across agent environments rather than rewriting skill content from scratch. Use it when skills silently stop activating, commands error on invocation, or IDE agent integrations regress after configuration edits.698installs35Repomix Safe Mixerrepomix-safe-mixer is a daymade/claude-code-skills utility for safe repomix runs. Before packing, it scans for hardcoded secrets—API keys, database credentials, tokens—reports findings, and removes or blocks unsafe content so distributed bundles do not leak credentials. Developers reach for it when creating reference packages for LLM context, open-source drops, or vendor handoffs where repomix is the packager of choice. The skill activates when users mention repomix packaging with security concerns or when sharing code externally. It adds a mandatory secret gate on top of repomix’s concat workflow, turning ad hoc packs into audited artifacts suitable for collaboration.680installs36Repomix Unmixerrepomix-unmixer is an agent skill from daymade/claude-code-skills that reverses Repomix packing by extracting individual files and restoring the original directory tree from XML, Markdown, or JSON single-file repository exports. Repomix consolidates entire repositories into AI-friendly archives, and repomix-unmixer unmixes those archives back into editable project layouts. Developers reach for repomix-unmixer when importing packed repos from collaborators, recovering structure from LLM context dumps, or continuing work on code that only exists inside a Repomix output file. Triggers include unmix, extract packed repository, or restore file structure from Repomix output.654installs37Financial Data Collectorfinancial-data-collector is a Claude Code skill that collects real financial data for any US publicly traded company from free public sources via yfinance. Output is structured JSON covering market data (price, shares, beta), historical income statements, cash flow, balance sheets, WACC inputs, and analyst estimates. Downstream skills consume the JSON for DCF modeling, comps analysis, and earnings review. Developers reach for financial-data-collector when a ticker symbol is given and clean inputs are needed before building valuation spreadsheets or automated finance agents—triggered by phrases like collect data for ticker, get financials for company, or gather DCF inputs.648installs38Product Analysisproduct-analysis is a daymade Claude Code skill that runs multi-path parallel product analysis using Claude Code agent teams plus Codex CLI for cross-model test-time compute scaling. Developers specify scope as full, ux, api, arch, or compare, and the skill explores the product from multiple AI perspectives before synthesizing findings into an actionable optimization plan. It can invoke competitors-analysis for competitive benchmarking and triggers on product audit, UX audit, information architecture review, or pre-release review requests. The core principle applies diverse model viewpoints to surface blind spots a single-agent pass misses. Teams reach for product-analysis before launches, after major refactors, or when internal self-review needs structured, cross-functional coverage across UX, API contracts, and system architecture.613installs39Doc To Markdowndoc-to-markdown is a documentation conversion skill from daymade/claude-code-skills that transforms Word documents, PDFs, and other rich office files into clean Markdown suitable for Git repositories, static docs sites, and agent-readable knowledge bases. Developers reach for doc-to-markdown when onboarding legacy specifications, compliance PDFs, or stakeholder Word drafts into a docs-as-code workflow without manually rebuilding every heading, list, and table. The skill focuses on structural fidelity—preserving headings, lists, and tabular data while stripping presentation cruft that breaks Markdown linters and static site generators.605installs40Windows Remote Desktop Connection Doctorwindows-remote-desktop-connection-doctor is a macOS-focused VDI troubleshooting skill from daymade/claude-code-skills for Microsoft Remote Desktop, Azure Virtual Desktop, and Windows 365 connections. It analyzes transport protocol selection between UDP Shortpath and WebSocket, detects VPN or proxy interference with STUN and TURN negotiation, parses Windows App logs for Shortpath failures, and explains unexpectedly high RTT. Developers invoke it when sessions feel slow, transport shows WebSocket instead of UDP, RDP Shortpath fails to establish, or latency spikes without an obvious network cause. Allowed tools include Read, Grep, and Bash for log inspection on the local Mac client.572installs41Scrapling Skillscrapling-skill is a Claude Code skill for building Python web scraping pipelines that survive DOM changes, anti-bot pages, and pagination when pulling external data into agents, APIs, or content systems. It guides agents to structure fetch-parse-store flows with defensive selectors and retry patterns instead of brittle one-off scripts. Developers reach for scrapling-skill when production ingestion must keep working after site layout shifts or bot challenges. Outputs are pipeline code and ingestion patterns rather than one-time HTML dumps.549installs42Continue Claude Workcontinue-claude-work is a daymade/claude-code-skills workflow that recovers actionable state from a prior Claude Code session and continues execution in the current conversation. Unlike claude --resume, which replays entire transcripts and wastes tokens on resolved threads, this skill selectively extracts the latest compact summary, pending todos, known errors, and current workspace state from local session files as the source of truth. A bundled extraction script handles session discovery, compact-boundary parsing, noise filtering, and workspace inspection in one call, documented in references/file_structure.md for JSONL schemas. Developers reach for continue-claude-work when a long Claude Code session stopped mid-feature and you need agents to pick up concrete edits and verification steps instead of re-explaining scope or re-reading the whole repository from scratch during multi-day agent workflows.547installs43Asr Transcribe To TextGuides building ASR transcribe-to-text capabilities inside Claude Code projects: capturing audio, invoking recognition services, normalizing output, and feeding transcripts to agents or databases. Covers error retries, chunking long recordings, confidence thresholds, and integration touchpoints for meeting bots, voice UIs, and support automation.542installs44Fixing Claude Export Conversationsfixing-claude-export-conversations is a Claude Code skill from daymade/claude-code-skills for salvaging Claude conversation export files that fail to import or load cleanly. The skill targets common JSON breakage—truncated message threads, malformed tool-use blocks, and schema inconsistencies—that block archival review, compliance audits, and dataset preparation for skill fine-tuning. Developers reach for fixing-claude-export-conversations when Claude Desktop or CLI exports stop parsing, conversation history appears cut off, or automated pipelines reject export JSON that must remain searchable and reproducible.540installs45Douban Skilldouban-skill is a Claude Code skill for integrating Douban media metadata into software projects. It helps developers retrieve book and film listings, ratings, and related catalog fields, then wire that data into content apps, chatbots, or research agents that reference Chinese-language media libraries. The skill supports automated publishing workflows that cite Douban sources correctly inside generated content. Use it when building recommendation features, media review bots, or catalog-backed research tools that need Douban as a data source. It assumes API or scraping patterns appropriate to Douban rather than generic REST boilerplate.511installs46Terraform Skillterraform-skill is a Claude Code skill from daymade/claude-code-skills that guides agents through Terraform for real cloud stacks. It covers writing modules for VPCs, IAM roles, RDS or managed databases, and attached services; enforcing remote state and workspace discipline; and running plan-before-apply review loops. The skill emphasizes module reuse, variable threading, output contracts, and change safety so infrastructure edits stay reviewable in pull requests. Reach for terraform-skill when you are standing up or extending AWS, GCP, or Azure resources and want Claude Code or Cursor to draft .tf files, catch anti-patterns, and walk terraform plan and terraform apply with explicit approval gates instead of one-shot HCL generation without operational guardrails. Review passes catch hard-coded secrets, missing backend blocks, and destructive replace operations before apply. The skill also threads outputs across nested modules so downstream services receive VPC IDs, subnet lists, and database connection strings consistently.478installs47Gangtise Copilotgangtise-copilot is a Claude Code skill for constructing finance copilots grounded in Gangtise market, company, and research data. It guides developers embedding those signals into analyst workflows, scheduled briefings, or internal chat agents that answer sector and issuer questions with sourced context. Engineers reach for gangtise-copilot when prototyping institutional research assistants, automating morning market summaries, or wiring proprietary terminals into agent toolchains. The skill assumes familiarity with Gangtise as a data source and focuses on copilot architecture rather than generic LLM prompting.474installs48Marketplace Devmarketplace-dev is a Claude Code skill from daymade/claude-code-skills that scaffolds complete two-sided marketplace backend architecture for MVP launches. The skill generates backend structures for product listings, search, shopping carts, seller onboarding flows, commission calculations, payout handling, and order state machines that model multi-party transactions. Developers reach for marketplace-dev when starting a marketplace SaaS or ecommerce API and need domain models, service boundaries, and transactional workflows rather than a simple storefront. The output targets production-minded MVPs where buyers, sellers, and platform fees interact through explicit state transitions and auditable order lifecycles.471installs49Tunnel Doctortunnel-doctor is a daymade/claude-code-skills diagnostic skill for macOS when Tailscale coexists with proxy/VPN tools like Shadowrocket, Clash, Surge, or OrbStack/Docker. It maps five conflict layers: route-table hijacking, HTTP_PROXY env vars, system proxy bypass, SSH ProxyCommand double tunneling, and VM/container proxy propagation—with step-by-step workflows for symptoms like Tailscale ping succeeding while SSH times out, browser 503s with working curl, git push HTTP relay failures, and 60-second DNS stalls in ssh -vvv. The skill uses Read, Grep, Edit, and Bash to verify each layer with authoritative health checks before applying fixes. Reach for tunnel-doctor when remote dev over Tailscale breaks under TUN proxies or Docker builds fail behind VPN.444installs50Debugging Network Issuesdebugging-network-issues is an evidence-driven methodology skill in daymade/claude-code-skills born from a documented five-hour production SSE case. It enforces eight steps: scope the symptom, verify the premise, distinguish upload- versus processing-timeouts on large POST bodies, gather per-hop evidence, frame falsifiable hypotheses, run layered isolation experiments, counter-review findings, then fix and re-run the same test. It covers ECONNRESET, HTTP/2 RST_STREAM, SSE stalls, Cloudflare 524/522 on uploads, client VPN/TUN misrouting, and Caddy or nginx `bytes_read` versus `Content-Length` splits. Domain skills like cloudflare-troubleshooting handle stack-specific tables; this skill supplies the general network discipline. Reach for debugging-network-issues when one log line feels insufficient. Skip it for pure application logic bugs with no network hop.427installs51Slides Creatorslides-creator is a narrative-first presentation skill in daymade/claude-code-skills. Phase 0 collects the user’s own transcripts, articles, and notes—AI invention without source material is explicitly forbidden. Phase 1 discusses story arc via the seven-step ABCDEFG framework (Attention through Go). Phase 2 structures machine-readable `content.md`; Phase 3 calls baoyu-slide-deck with `--prompts-only` then `--images-only` for outline, images, PPTX, and PDF. Phase 6 reorganizes output and extracts speaker notes into `00-上游/`. The skill forks a general-purpose agent context and triggers on “create slides”, “PPT”, or “slide deck”. Reach for slides-creator when turning existing writing into a defensible talk in the user’s voice. Skip it for quick bullet dumps without source material or when baoyu-slide-deck is unavailable.424installs52Competitors Analysiscompetitors-analysis is a competitor research and positioning skill that helps developers and product engineers turn messy market signals into an actionable competitive map. competitors-analysis structures rival analysis around positioning, pricing tiers, feature gaps, and go-to-market motions, so teams can make clearer decisions about what to build and how to package it. Developers typically reach for competitors-analysis when requirements feel ambiguous, pricing or packaging is unsettled, or a roadmap needs sharper differentiation to reduce wasted engineering effort. competitors-analysis is also useful when preparing a narrative for product strategy docs, planning experiments, or aligning stakeholders around a defensible scope based on competitor coverage and GTM patterns.417installs53Stepfun Ttsstepfun-tts is a daymade/claude-code-skills integration for StepFun stepaudio-2.5-tts, released 2026-04 and verified 2026-04-23. The skill replaces step-tts-2 voice_label tags with natural-language instruction fields up to 200 characters plus inline () prosody directives inside the spoken text. Bundled scripts include tts_generate.py for single-line or JSONL batch synthesis, ab_compare.sh for model A/B checks, and three reference guides covering API JSON, migration from step-tts-2, and censorship pitfalls. Developers reach for stepfun-tts when building game voice lines, accessibility narration, podcast drafts, or agent responses that need whisper, pause, stress, or mood control without maintaining a separate audio toolchain. The workflow handles censorship_block responses per line, enforces a 1000-character hard input cap, and stores API keys in STEPFUN_API_KEY or CLAUDE_PLUGIN_DATA config.json. Pricing guidance in the skill cites roughly 5.8 yuan per 10,000 characters for contextual synthesis.412installs54Feishu Doc Scraperfeishu-doc-scraper is a Claude Code skill for extracting structured text and metadata from Feishu/Lark (Lark Suite) documents into developer workflows. When product specs, API contracts, or operational runbooks live in enterprise collaboration tools instead of git, this skill retrieves readable content for repos, issue trackers, or retrieval-augmented generation corpora. Developers reach for it during onboarding to internal platforms, when syncing a Feishu PRD into docs/, or when agents need grounded context from locked collaboration spaces. The skill targets the gap between walled-garden documentation and code-adjacent markdown engineers and CI can index. Use it before implementing features documented only in Feishu or when building RAG pipelines that must include Lark-hosted knowledge bases.381installs55Stepfun Asrstepfun-asr teaches Claude how to call StepFun automatic speech recognition: prepare audio formats, manage API keys, poll or stream results, and attach transcripts to downstream LLM prompts. It supports voice-first agents, meeting capture, and accessibility features that need accurate Mandarin or multilingual transcription inside production codepaths.361installs56Benchmark Due DiligenceA developer tool for AI integration and automation. This is a developer tool for building and integrating AI-powered features.248installs57Bigdata SkillA developer tool for AI integration and automation. This is a developer tool for building and integrating AI-powered features.246installs58Auto Repo SetupAutomates codebase environment setup, dependency diagnosis, and safe git operations by reading ONBOARDING.md and validating the project runs. A developer or non-technical user uses it when a cloned repo won't start or standard setup is needed.240installs59Pdf To HtmlConverts a PDF into a single self-contained HTML file preserving images, tables, charts, and reading order, with optional translation. A developer uses it when they want to read or translate a PDF as a faithful web page.211installs60Terminal ScreenshotRenders a terminal command's ANSI-colored output to a PNG image so the visual result can be judged instead of reading raw escape codes. A developer uses it to verify CLI color config like delta, bat, or theme changes.200installs61Llm Wiki SetupHelps build a personal investment-research LLM Wiki (Karpathy pattern) as pure markdown plus wikilinks maintained by the LLM, by interviewing the user rather than handing a template. A developer uses it when building a compounding research knowledge base.196installs62Bilibili SourceFetches real, verifiable Bilibili video metadata and engagement stats via a single API call. A developer uses it when ingesting a Bilibili source or citing accurate view/like counts instead of estimating.170installs63Claude Usage AnalystProduces evidence-based explanations of Claude Code token usage, cost, and quota burn using local ccusage data. A developer uses it when investigating why Claude quota was exhausted or which model is expensive.170installs64Marketplace Health CheckRuns a six-dimension parallel health check of a Claude Code skills marketplace repo covering code safety, docs consistency, PII, PR/issue triage, and manifest integrity. A developer uses it to audit the whole repo before a release.164installs