Heidelberg AI News

Video · GitHub ·

How a personal AI productivity hack became Slack's internal tool

In short

Frances Coronel explains how a personal Claude Code setup evolved into Slack’s DevX PI catalog, a centralized platform for AI tools, MCP servers, skills, and coding agents. The catalog combines self-service access with telemetry, enabling Slack to measure adoption, guide experimentation, and spread AI practices across the organization.

Key points

  • The setup grew from 10 engineers copying Frances’s shared configuration into a catalog with 300+ agentic tools.
  • The catalog tracks 60 MCP servers built by 33 teams and more than 4 million MCP tool calls.
  • Tool inventory grew from 85 in March to 310, a 233% increase.
  • GitHub Actions synchronize catalog data daily, while an MCP server makes it queryable by Slackbot, Spec, Claude Code, Codex, and Cursor.
  • AI agents automate pull-request creation, review, ownership checks, and approvals; the pattern has supported 3,500+ PRs across 15 repositories, about 1,500 per week.
  • Slack’s AI culture expanded to 1,000+ people in a Claude Code channel, and 90% of surveyed employees reported actively sharing AI prompts and tips.
  • The recommended approach is to ship an imperfect prototype publicly, centralize tools, measure usage as a product, and encourage experimentation.

Written by an AI model from the video’s captions and description.

Frances Coronel, a product engineer at Slack, built a personal AI setup to work faster and shared it in a public channel. Engineers across the company started forking it without being asked.

Watch on YouTube The opening of the article at youtube.com.

More from GitHub

All stories

VideoGitHub

The view from the labs: What's next for AI-assisted development

يناقش panel من Anthropic وOpenAI وGoogle Cloud وGitHub كيف تنتقل البرمجة بالذكاء الاصطناعي من المساعدة المباشرة إلى وكلاء مستقلين وأنظمة وكلاء تنفذ العمل بالكامل. تكمن القيمة القادمة في بناء أنظمة قابلة للتحقق، وتأمين هوياتها وصلاحياتها، وامتلاك الحكم البشري لتحديد ما ينبغي بناؤه ولماذا ومتى.

VideoGitHub

LangChain Deep Agents

The video introduces LangChain Deep Agents, an open-source framework that provides opinionated defaults and best practices for building agents with any model. It covers sub-agents, skills, memory, proactive behavior, decision models, safeguards, and cloud execution, explaining why these primitives matter for reliable long-running and personal agents.