Workshop repo: when to use Claude Skills vs MCP tools vs agent definitions
cfregly · x · 2026-11-01
cfregly's learn-claude-workshop repo offers a practical chooser for Skill/Tool/MCP/agent-definition layers, demoed on a fictional Northstar Freight renewal-brief task. Key tradeoffs: custom Skills reuse know-how but need upkeep; MCP connects tools but needs access and failure handling; agent definitions reuse worker settings but still need review; sometimes a plain script suffices. The repo includes an MCP server with denial checks, a Skills API bundle pinned by version, a bundled self-checking renderer, and a local evidence checker comparing tool results, file hashes and outputs.
Related event: Claude Workshop Offers Layered Guidance on Choosing Skills, MCP and Agents(2 posts)→
More from coding & agent
- One Claude Code command audits your Managed Agents setup against Anthropic's security guide — brada · 2026-11-01
- Claude-Code-Game-Studios: open-source template turns Claude Code into a 49-agent game studio — lxfater · 2026-11-01
- OpenMoat: open-source local policy checks to secure AI coding agents with machine access — __dotk__ · 2026-11-01
- Senior engineer: 'popularity is the weights of an LLM' — half of 600 engineering scars can't be baked into models — RandalSchwartz · 2026-11-01
- Factory's agent ships an MCP connector registry; david hoang wants Paper added — davidhoang · 2026-11-01
- Stardock's GalCiv IV dev shows how AI agents kill 5:1 asset-pipeline overhead and finally clear years of player requests — draginol · 2026-11-01