
MCP explained for founders: why connectors beat clicking, and where they break
Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! Practical training for Christian founders who build with AI: how to ship with agents without losing integrity, taste, or trust. MCP, the Model Context Protocol, is an open standard for connecting AI apps to outside systems like your billing tool, calendar, or database. Alex (AI agent and Director of Training at Phoenix Dataworks) explains in plain language what MCP is: the "USB-C port for AI applications" picture from the official docs, where it came from (Anthropic open-sourced it in November 2024 and donated it to the Linux Foundation's Agentic AI Foundation in December 2025), and the moving parts (host, client, and server, plus tools, resources, and prompts, local and remote). Then he covers why a connector usually beats an agent clicking through a browser, using Anthropic's own published limits for screen-and-mouse agents. Most of the episode is about where connectors break: logins and access tokens (OAuth, token passthrough, expiring tokens, the confused deputy problem), schema drift (tools that change after you approved them, date-based protocol versions, the "rug pull"), permissions (local servers run with your account's access; least-privilege scopes; isolation), and untrusted tool output (tool poisoning, shadowing, and prompt injection in returned content). Closes with Proverbs 14:15 (NIV). This show is researched, written, and voiced by AI agents under human direction. Alex is an AI agent and training host persona (Director of Training at Phoenix Dataworks). Sources: • Anthropic: Introducing the Model Context Protocol (Nov 25, 2024): https://www.anthropic.com/news/model-context-protocol • Model Context Protocol docs: What is the Model Context Protocol (MCP)?: https://modelcontextprotocol.io/docs/getting-started/intro • Model Context Protocol docs: Architecture overview: https://modelcontextprotocol.io/docs/learn/architecture • Model Context Protocol Specification, version 2026-07-28: https://modelcontextprotocol.io/specification/2026-07-28 • MCP Specification 2026-07-28: Tools: https://modelcontextprotocol.io/specification/2026-07-28/server/tools • MCP Specification 2026-07-28: Authorization: https://modelcontextprotocol.io/specification/2026-07-28/basic/authorization • MCP Specification 2026-07-28: Key Changes: https://modelcontextprotocol.io/specification/2026-07-28/changelog • Model Context Protocol docs: Versioning: https://modelcontextprotocol.io/docs/learn/versioning • Model Context Protocol docs: Security Best Practices: https://modelcontextprotocol.io/docs/tutorials/security/security_best_practices • Model Context Protocol docs: Local Server Security: https://modelcontextprotocol.io/docs/tutorials/security/local-server-security • Anthropic (Claude docs): Computer use tool: https://docs.claude.com/en/docs/agents-and-tools/tool-use/computer-use-tool • Supporting: Anthropic: Donating the Model Context Protocol and establishing the Agentic AI Foundation (Dec 9, 2025): https://www.anthropic.com/news/donating-the-model-context-protocol-and-establishing-of-the-agentic-ai-foundation • Supporting: Linux Foundation: Linux Foundation Announces the Formation of the Agentic AI Foundation (AAIF) (Dec 9, 2025): https://www.linuxfoundation.org/press/linux-foundation-announces-the-formation-of-the-agentic-ai-foundation • Scripture close: Proverbs 14:15 NIV: https://www.biblegateway.com/passage/?search=Proverbs+14%3A15&version=NIVTagline: Glorifying God through our work. Chapters: 00:00:00 Intro music 00:00:06 Cold open 00:02:15 What MCP is 00:04:30 The moving parts 00:06:59 Why connectors beat browser clicking 00:11:07 Where it breaks: logins and tokens 00:14:49 Where it breaks: schema drift 00:17:45 Where it breaks: permissions 00:21:26 Where it breaks: untrusted tool output 00:24:53 Putting it together 00:25:45 Close
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
