Weekly Newsletter Digest
Read time: 2 min 14 sec
- News: Original pieces for AI Native developers, and 6 items on our radar.
- DevCon: Join us in NYC (or online!)
- Landscape: 6 new tools to highlight.
- Podcast: Redefining developer workflows in the AI era with MCP.
1. News
GitHub's Spec Kit structures intent for AI agents: Markdown-based templates turn project requirements into version-controlled blueprints, replacing mind-reading prompts with explicit specifications.
Rulebook-AI compiles shared context into assistant configs: Open-source CLI turns unified rulesets into tool-specific files, stopping fragmentation across Cursor, Copilot, and Claude Code.
Codeplain goes all-in on spec-driven development: Plain language turns structured specs into consistent, production-ready code, solving AI's 'functionality flickering' beyond vibe-coding.
GitHub launches curated MCP registry for AI interoperability: Repository-backed listings bring trust signals to MCP servers, solving discovery as agents wire into external tools.
AI Native Dev Radar 📡
- OpenAI introduces AgentKit
- Meet Jules Tools, Google’s CLI Companion
- Building consistent workflows with Codex Agents SDK
- Is 90% of code going to be written by AI?
- Directing Swarms in the AI Era
- Code Mode: The better way to use MCP?
2. Join us at DevCon Fall
Early bird tickets now available
On November 18–19, DevCon arrives in Brooklyn, New York! Our worldwide community of AI-native builders will meet face-to-face (and online) to explore spec-driven creation, coding agents, breakthrough dev tools, and real-world AI development insights from the front lines.
3. Landscape
Your Guide to the AI Development Ecosystem
This week brought a wave of new tools, and we’re highlighting 6:
- Conare, context management for Claude Code.
- Supermemory, universal memory API for AI apps.
- Codex-Action, run Codex from a GitHub Actions workflow.
- Traycer, spec-driven AI generation with planning first.
- StackBench, tests agent-readiness of your docs.
- Userdoc, reverse-engineers code into AI-ready specs.
Discover more AI dev tools on Landscape.
4. New on the podcast
Redefining Developer Workflows in the AI Era with MCP
The risk of letting AI do more than autocomplete? It can quickly spin out of control. On this episode of AI Native Dev, Steve Manuel, founder and CEO of Dylibso, unpacks MCP, the protocol that keeps AI extensions safe and predictable, and dives into mcp.run, his framework for tapping into shared MCP servers without losing control.