Recently Evaluated

A Zenodo preprint reporting a 12,160-trial black-box evaluation of GPT-5.4 outputs under two prompt conditions, with detailed token-ceiling sweeps, control/ablation trials, and hash-chained verification.

Open-source, self-hostable platform for shipping Claude Code-style AI agents to Kubernetes or local Docker Compose, with Slack as the first supported channel. The README shows a command-line workflow, deployment tiers, and support for traces, evals, budgets, and model choices.

Kotro is a local control plane for coding agents that governs MCP tool actions and LLM traffic with one binary. The page shows install options, supported clients, dashboard access, and concrete controls like schema pinning, redaction, cache, budget, and circuit breaker behavior.

A CLI for running untrusted MCP servers in sandboxed containers with deny-by-default egress, traffic watching, and host approval workflows. It can also compose servers into agents and expose them over OCI registries.

clai is a terminal-first AI tool that pipes stdin to an LLM and writes results to stdout, so it fits naturally into shell workflows like git diffs, text files, and command pipelines.

FastAPI-based server that turns files and Python modules into REST APIs, browsable UIs, searchable resources, and an MCP endpoint.

An AI-native 3D editor built on Blender with an embedded agent for 3D production tasks like retopology, UV rebuilds, baking, and export prep.

Bullet is a fast coding agent for macOS that focuses on lower-latency routing, targeted code search, and parallel tool execution. The page highlights a private beta, free access, and an internal demo with SWE-bench Verified results.

Compass is a local-first code knowledge graph engine in Rust that turns codebases and project artifacts into searchable, queryable graphs. The page shows support for architecture mapping, change-impact analysis, historical graph diffs, deterministic CompassQL queries, and integrations like VS Code, MCP, and export form

A Vite plugin that obfuscates client-side code in a way meant to make LLMs refuse de-obfuscation attempts. The README shows the project’s purpose, a usage note, and an example of the transformed output.

An architecture case study showing how a Slack-like app is built in pure Elixir with Jido agents, messaging, chat bridges, and Hologram UI.

A standalone CLI coding agent from Warp that runs in any terminal and supports multi-model routing, persistent sessions, full-screen terminal apps, and cloud-agent handoff.

A Hacker News reader focused on searching stories and comments, filtering noise, and browsing discussions with recency and text-based matching controls.

Stonefold is a deterministic enforcement gateway for AI agents. It sits between an agent and real systems, validates typed intents, applies policy, supports human approval and kill-switch controls, and records an audit trail for every attempt.

Macro is an open-source team workspace that combines email, chat, docs, tasks, agents, calls, and CRM in one linked system with shared AI memory.

Research preview for a voice-agent reasoning harness that lets a realtime assistant think between turns without adding in-band latency. The page includes benchmark results, pricing, example behavior, and a description of the architecture behind Waterr’s AI meetings.

Open-source security guidance for specific libraries and versions, aimed at developers and coding agents. The repo includes local setup steps, a web site, and an installable skill for agent workflows.

An open-source app that takes a finger-frame gesture video and turns the framed area into an AI-generated scene, then composites it back into the original clip. It includes a browser demo, local run instructions, and Python CLI scripts for offline use.

A Vale style pack that lints prose for AI-like writing, inspired by ASD-STE100. It helps teams catch buzzwords, vague phrasing, self-praise, and other patterns in docs, comments, and agent-authored text.

Local-first private planning and decision workspace that turns messy documents into structured plans, grounded outputs, and traceable decisions with citations and audit trails.

A Git-backed CLI for sharing agent skills, settings, and MCP fragments across repositories with per-project lockfiles and update workflows.

An open-source agent-driven package manager for versioned, checksummed, approval-gated tasks run through Claude Code, Codex, Gemini CLI, OpenCode, or Antigravity.

Terminal-native attention router for parallel Claude Code sessions. It shows session status in the terminal, highlights the one that needs help, and provides quick jump shortcuts, tmux support, and optional desktop notifications.

crosswalk is an MCP-based social context-sharing product for coding agents, letting teams share searchable posts and pull group knowledge into tools like Claude Code, Cursor, Codex, and VS Code.