GitHub Repo Briefing
strix
An AI-assisted penetration-testing tool that scans a local application, explains possible vulnerabilities, and helps investigate fixes.
A first test
Run strix --target against a deliberately vulnerable local app. Reproduce one finding manually, record the request and fix, and use it only on systems you are authorised to test.
ai-berkshire
A set of value-investing research skills for Claude and Codex that combines financial statements, valuation methods, and adversarial review into a report.
A first test
Run one skill on a public annual report and keep the source pages with the output. Recalculate one key ratio yourself and treat the result as research, not a recommendation.
simplex-chat
A private messaging system whose network avoids global user identifiers and provides mobile, desktop, and terminal clients.
A first test
Install two test clients and send a message over a fresh connection. Check identity setup, backups, notifications, and metadata behavior before inviting a real contact.
lingbot-map
A 3D streaming-reconstruction model that builds a spatial map from visual input for research and interactive scene understanding.
A first test
Create the documented environment and run the demo on a short local sequence. Compare the reconstructed points with the input frames before using it for measurement.
OmniRoute
A local AI gateway that exposes one endpoint for many model providers and translates requests for clients such as Claude Code, Codex, Cursor, and Copilot.
A first test
Run it on the local port with one provider and one low-risk prompt. Inspect the selected model, request log, costs, and fallback behavior before adding credentials for other providers.
codebase-memory-mcp
A local MCP server that builds code intelligence from Tree-sitter parsing and a graph, with a static binary and support for many programming languages.
A first test
Install the binary and index a small repository. Ask for a symbol definition and callers, compare both with the source, and check the MCP client permissions before enabling edits.
herdr
A terminal agent multiplexer that runs several coding agents in parallel and provides a shared control surface for their sessions.
A first test
Run it in a disposable repository with two read-only tasks. Check each session's working directory and process before allowing one agent to edit or commit files.
logto
An open-source identity service with OIDC and OAuth 2.1, multi-tenancy, single sign-on, role-based access, and user management.
A first test
Start the Docker Compose example and create a test tenant and application. Verify login, token claims, redirect URLs, and role checks before connecting a production client.
OpenSuperWhisper
A macOS dictation application that runs Whisper or Parakeet speech models locally and inserts the transcription into the active application.
A first test
Install it with brew install opensuperwhisper and download one local model. Dictate a short note, check punctuation and hotkeys, and confirm the audio stays local.
video-use
A coding-agent tool for editing video from a footage folder and a natural-language request, with a final video written to the project directory.
A first test
Give it three short test clips and ask for one cut. Inspect the edit decision, timestamps, and final.mp4, then confirm that the source files were not overwritten.
agency-agents
A collection of specialised agent roles with defined responsibilities, processes, and deliverables for engineering, research, marketing, and operations work.
A first test
Choose one role and give it a bounded task with a required output format. Check its assumptions and sources, then combine it with a human review instead of treating the role description as expertise.
OpenMontage
An agentic video-production system that coordinates media preparation, planning, editing, and rendering through a pipeline called Backlot.
A first test
Run make setup and open one sample project with python -m backlot open. Use test footage, inspect each pipeline step, and verify the final render before adding client media.