The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Created a dashboard for a fashion brand.
Running a small fashion brand means checking every number yourself. With an AI agent that runs the checks, this make management much easier.
Analytics: ask what needs attention and get a ranked list.
Sals: click a day on the chart and the whole page follows.
History: every question is saved with a snapshot of the page.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
An open-source MCP (Model Context Protocol) server that lets AI assistants like Claude work with X directly, using 11 tools built on the X API v1 and v2.
Agents can post and search content, look up and update profiles, follow and unfollow accounts, list followers and following, and create and manage public or private lists.
It's set up with environment-based API keys and stdio transport, plus GitHub Actions for publishing and releases. It's part of my open-source MCP series, alongside servers for LinkedIn, Medium, and Kaggle.
You've built MCP servers for a few different platforms now. Have you found any tool names or descriptions that made a noticeable difference in how reliably the agent chose the right tool? I'd be interested to hear what you ended up changing.
Where should an agent’s permission check live?
I built DeadBolt at NextEleven to put that check at the point where a tool action actually runs, outside the model’s instructions. Stable 2.0.0 is now available, free and open source under Apache-2.0.
For example: an agent requests a file write. Your executor asks DeadBolt first. A denial leaves the protected function uncalled. You can set tool policies, require one-use approvals, pause a run, or kill its future admissions. Store or transport failure denies protected dispatch.
You can embed it in Rust, use a local sidecar from Python or another language, or wrap a trusted stdio MCP server.
The boundary matters: DeadBolt protects routes you connect. It is not a sandbox for arbitrary code, and kill does not automatically cancel work already running.
Install, examples and captioned walkthroughs:
https://nexteleven.dev/deadbolt
If you’re building agents, what would make a permission gate easier to integrate into your stack? I’m happy to answer setup questions and consider useful feedback.
nexteleven.dev
DeadBolt — Give agents tools. Keep control.
A local execution gate for software agents, built by NextEleven.
For me, a small MCP example would help: approve a file write, then show what happens if the agent changes the path or content before execution. Is the one-use approval tied to those exact arguments?
InteractPro- Realtime Collaboration & AI Chat Platform
Live link: https://interactpro.vercel.app/
Github link: https://github.com/nitinjha7/InteractPro-ChatApplication
React, TypeScript, Node.js, Express, tRPC, PostgreSQL, Prisma, Socket.IO, Yjs, Gemini, pgvector
A developer-focused platform combining real-time messaging, collaborative code editing, persistent sessions, and AI-powered semantic search over private conversations.
Key features
1) Real-time one-to-one chat and syntax-highlighted code sharing.
2) Concurrent collaborative editing using Yjs CRDTs and WebSockets.
3) Persistent code sessions with database-backed state recovery.
4) RAG-based semantic search using Gemini embeddings, PostgreSQL pgvector and HNSW indexing.
5) Answers grounded in relevant messages with citations.
6) 71 manually written server-side tests.
What was the trickiest part of getting Yjs and the saved sessions to work together? I'm curious how you handled someone reconnecting after others had kept editing.