Cracking Shells

It came out prettier than I thought! 😊 Just released Hatch v0.8.0. This update is all about UX: clean tables, helpful colors, outputs that actually make sense. 🐣 `pip install hatch-xclam` #MCP #AIForScience

*Hatch!* v0.7.1 is out! 🐣 github.com/CrackingShel... It adds support for managing #MCP servers in - Codex - Kiro #AIForScience

Name tags for our #MCP hackathon next week! Embracing Japan's 紅葉 season🍁🍂 for the theme 🐣 #AIForScience

Just stumbled upon Augmented Nature's #MCP server collection for #LifeSciences. UniProt, AlphaFold, KEGG, PDB (and so many more!), all exposed via MCP: 🔗 github.com/Augmented-Na... #AIForScience #Biology

Solid #MCP design patterns in @anthropic.com ’s latest: "Code execution with MCP: Building more efficient agents" 🔗 www.anthropic.com/engineering/... Exactly the kind of resource to share with your coding agent before tackling a server. I appreciate the didactic effort. 🐣 #AIForScience

Two #MCP servers on Docker Hub for scientific papers 📚 🔗 paper-search: hub.docker.com/r/mcp/paper-search 🔗 arxiv-mcp-server: hub.docker.com/r/mcp/arxiv-mcp-server Useful for automating literature discovery in #AIForScience workflows. Have you tried them yet? #MCP #OpenScience

BioMCP: 24 #MCP tools for #biology 🧬 Shoutout to @genomoncology.bsky.social ! Search PubMed, ClinicalTrials, cBioPortal...Fetch genes/variants, expand disease synonyms, match trials to biomarkers, analyze OpenFDA safety data... 🔗 biomcp.org (Had forgotten about that one🤦) #AIForScience

Last week's #DevLog Hatch! 0.7.0 now lets you add MCP packages with one command—It's multi-dependencies and multi-host. Finally some versatility for #MCP setup 🐣 Still a pre-release, bug fixes & docs in progress! #AIForScience

AI Jam incoming in Japan! trip.riken.jp/en/aijam/ind... A hands-on event where scientists collaborate with LLMs from AWS, Anthropic, Google, and OpenAI to tackle challenging #research problems. 📆Dec 16-18: Main Session (Tokyo/Kobe) with on-site support from AI experts. #AIForScience

For #AIForScience stack, #MCP servers for databases are crucial. A couple more of well-documented ones to explore: 🔗 Graph-based (for knowledge graphs): ArangoDB github.com/PCfVW/mcp-ar... 🔗 Document-based (for descriptive data): MongoDB github.com/mcp/mongodb-...

GitHub just launched the MCP Registry! It's the new place to find MCP Servers for any AI tool that speaks MCP, including GitHub Copilot and agents. Find what you need faster at: github.blog/ai-and-ml/gi... #GitHub #MCP #Developer #AI #OpenSource

Great to see the MCP tool scalability problem getting attention! As the article notes, it’s a critical bottleneck, but solutions are coming. We’re tackling this too at Cracking Shells (ETA: December). The more minds on it, the faster we’ll crack it. 🐣 #MCP #AIForScience

Using the server itself as a layer helping to manage the tools is indeed within the benefit of #MCP we'd cite. But the top one for us remains the encapsulation and being able to (almost) drop it anywhere. This is so convenient to decouple dev work from actual agent systems, particularly in #science

Three #MCP servers for mechanistic modeling in #biology. A well-rounded example of making LLMs reliable for scientific applications. #AIForScience www.biorxiv.org/content/10.1...

After reading the MCP Registry docs, a few thoughts: 1. Not a perfect fit for #MCP for #science (yet), but that’s okay because 👇 2. Recommend filtering of existing registries AND 3. Support for custom package registries: the whole point of Cracking Shells 🤩—MCP servers for scientific applications

The MCP Registry is live! 🎉 Huge for tool discovery—especially if it works well for #MCP in #science. At Cracking Shells, we’ve built our own solutions, but we’re all for interoperability. Excited to test, compare, and adapt to what serves the community best. #AIForScience

For the open-source and decentralization folks: This session at the MCP Dev Summit is for you. Tadas Antanavicius and Toby Padilla are discussing the creation of an MCP Registry. A key step for the ecosystem. London, Oct 2. Register at mcpdevsummit.ai #OpenSource #Decentralization #AI #TechTalks

For the tech and AI folks on Bluesky: Check out this talk from Andrew Culberson at the MCP Dev Summit. He's covering how MCP is the backbone for multi-agent, multi-user workspaces. A key topic for anyone building collaborative tools. See you in London on Oct 2. #TechTalks #AIagents #LondonTech #MCP

Last week, we hosted our first #MCP hackathon for science with 22 #researchers (computational physics, chemistry, biology). None had used MCP or LLMs APIs before. By the end, they’d created their own MCP servers. This was invite-only, but more events are coming! 🐣 #AIForScience

Security is non-negotiable. At the MCP Dev Summit in London, Vitor Balocco will present "MCP Security: The Exploit Playbook." A crucial session covering real-world exploits and defensive strategies for the age of autonomous agents. 🗓️ Oct 2|London mcpdevsummit.ai #MCPDevSummit #AISecurity #MCP #Ai

#MCP isn’t the only way to connect LLMs to tools, but it won us over for #research: Small: the decorator pattern of the Python SDK is a killer to wrap existing API Decoupled: Devs build tools; users install them like packages (almost). Fits the ‘publish once, use anywhere’ model scientists like.

Cracking Shells is an open-source project to help #researchers integrate their scientific software with LLMs using the Model Context Protocol #MCP. We are building it because we need it, and maybe you do too 😉 We'll be sharing more about it. In the meantime, take a look: github.com/CrackingShells