Loomscope
Password-protected dashboard and Remote MCP observer for Codex work, recording concise task milestones via observe_start, observe_report, and observe_finish tools.
README
Loomscope
Agent work, visible. Loomscope is a password-protected dashboard and a Remote MCP observer for long-running Codex work. It records concise task milestones rather than private reasoning or raw logs.
Why one deployment
This repository is intentionally a single Cloudflare Worker deployment:
src/web/is the Vite + React + Tailwind dashboard.src/worker/is the independently testable HTTP, auth, MCP, validation, and storage boundary.migrations/owns D1 schema changes.plugins/loomscope-observer/contains the installable Codex plugin: its reporting Skill and Remote MCP configuration.
The UI and API share an origin, so no CORS policy or second deployment needs to be maintained. Static assets are served by the Worker, while /api/* and /mcp remain dynamic.
Setup
- Install dependencies:
pnpm install. - Copy
.dev.vars.exampleto.dev.varsand set three distinct long secrets. - Authenticate Wrangler:
pnpm exec wrangler login. - Apply schema:
pnpm exec wrangler d1 migrations apply loomscope --remote. - Check locally:
pnpm build, thenpnpm dev:worker. - Deploy:
pnpm deploy.
wrangler.jsonc is already bound to the production loomscope D1 database. If you intentionally deploy this project to a different Cloudflare account, create a new D1 database and replace that binding's database_id first.
The deployment configuration provisions loomscope.vortexislabs.com as the Worker custom domain. Before the first production deploy, set production secrets without putting them in wrangler.jsonc:
pnpm exec wrangler secret put DASHBOARD_PASSWORD
pnpm exec wrangler secret put DASHBOARD_SESSION_SECRET
pnpm exec wrangler secret put MCP_WRITE_TOKEN
Codex integration
The plugins/loomscope-observer/ directory is the source for the loomscope-observer Codex plugin. It contains the reporting Skill and the deployed Remote MCP endpoint. Start a new Codex thread after installation so its Skill and MCP tools are discovered.
The plugin reads its write token from LOOMSCOPE_MCP_TOKEN. For Codex CLI, set the variable before starting Codex:
export LOOMSCOPE_MCP_TOKEN="your-agent-write-token"
codex
The Worker exposes observe_start, observe_report, and observe_finish. The dashboard only accepts its own signed, password-created session; the MCP token has no read access.
Commands
pnpm lint
pnpm test
pnpm build
pnpm check
pnpm dev:worker
pnpm deploy
Data boundaries
D1 stores normalized projects, agents, runs, and event summaries. Large artifacts are represented by links in the initial release; add R2 only when artifact upload becomes a product requirement. This keeps reporting fast and avoids turning the observability database into an unbounded log store.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
E2B
Using MCP to run code via e2b.