daymade/claude-code-skills

• **Archetype**: mega-skill-pack
• **Stars**: 822
• **Last pushed**: 2026-04-12
• **License**: MIT
• **Discovered**: 2026-04-12
• **Skills found**: 51

Summary

Massive collection of 51 Claude Code skills organized as a marketplace. Covers an extraordinarily diverse range: research (deep-research with lead agent + subagent architecture, competitors-analysis, product-analysis), development (iOS-APP-developer, terraform, cloudflare-troubleshooting), media (ASR transcribe, video-comparer, youtube-downloader, capture-screen), productivity (excel-automation, i18n-expert, github-ops, github-contributor), documentation (suites/daymade-docs/ with 6 doc skills), content (wechat-article-scraper, twitter-reader, teams-channel-post-writer), DevOps (tunnel-doctor, windows-remote-desktop-connection-doctor), and meta-skills (skill-creator, skill-reviewer, skills-search, marketplace-dev). The deep-research skill is standout -- a multi-agent architecture with lead coordinator + parallel subagents for context-efficient research.

Assessment

HIGH VALUE. The deep-research skill alone justifies high priority: it implements a lead agent + subagent architecture where subagents write distilled notes to files, achieving ~60-70% context reduction. The six-phase pipeline (environment setup -> task board -> parallel dispatch -> citation registry -> evidence-mapped outline -> counter-review -> verify -> polish) is a sophisticated methodology. The source governance model (source types, freshness checks, circular verification forbidden) adds rigor. Beyond deep-research, the collection demonstrates remarkable breadth across domains. The daymade-docs suite (doc-to-markdown, docs-cleaner, meeting-minutes, mermaid-tools, pdf-creator, ppt-creator) is a cohesive document workflow. The meta-skills (skill-creator, skill-reviewer, skills-search) form a self-sustaining skill ecosystem.

Extraction Priority

• High
• Rationale: deep-research is a sophisticated multi-agent research methodology directly extractable as a babysitter process. The lead agent + subagent pattern with file-based context passing maps perfectly to babysitter's orchestration model. The document suite and meta-skills provide additional extraction targets.

Processes

• **Deep Research Process**: Setup source policy -> build research task board with roles/queries/parallel groups -> dispatch subagents in parallel -> build citation registry with source_type + as_of + authority -> create evidence-mapped outline with counter-claim flags -> draft from notes (never from raw search) -> counter-review (claims, confidence, alternatives) -> verify traceability -> polish with confidence markers. A 7-phase babysitter process with parallel subagent dispatch.
• **Document Conversion Suite**: 6 interconnected document workflows (Markdown conversion, cleaning, meeting minutes, Mermaid diagrams, PDF creation, PPT creation) that could be a babysitter process library under specializations/shared/document-processing.
• **Skill Development Lifecycle**: skill-creator -> skill-reviewer -> skills-search -> marketplace-dev. A meta-process for creating, evaluating, and distributing skills.
• **Competitive Analysis Process**: Research competitors -> analyze product positioning -> compare features -> generate report. Extractable to specializations/business/competitive-analysis.

Plugin Ideas

• **Deep Research plugin**: Install.md-driven plugin implementing the lead agent + subagent research architecture with source governance, citation registry, and counter-review. Configurable for enterprise research, general research, standard, or lightweight modes.
• **Document Processing Suite plugin**: Bundle of 6 document skills (Markdown, cleaning, meeting minutes, diagrams, PDF, PPT) as a cohesive document workflow plugin.
• **Skill Ecosystem Toolkit plugin**: Meta-skills for creating, reviewing, searching, and publishing skills. A self-bootstrapping skill development environment.

Patterns

• **Lead agent + subagent architecture**: Coordinator agent dispatches parallel subagents who write distilled notes to files. Lead agent synthesizes from notes only, never from raw search results. ~60-70% context reduction.
• **Source governance model**: Source types, as_of freshness checks, authority levels, circular verification forbidden, exclusive advantage encouraged. Adds rigor to research processes.
• **Citation registry**: Structured tracking of every source with type, freshness date, authority level. Enables traceability verification in final output.
• **Counter-review phase**: Mandatory phase that challenges claims, assesses confidence levels, and identifies alternatives. Prevents confirmation bias in research.
• **Mode selection matrix**: Two dimensions (topic mode x depth mode) producing four configurations. Clean parameterization of process behavior.
• **Document workflow suite**: Cohesive set of related skills (suites/daymade-docs/) that share context and can be chained. A skill-suite organizational pattern.

Library Mapping

Plugin Marketplace Mapping