Recently Evaluated

An open 30B agentic model from Meta designed to run locally on consumer hardware for coding, function calling, and other always-on agent workflows.

A Microsoft Research paper PDF on agentic coding and GitHub Copilot at production scale. The crawl confirms it is a substantive research document tied to an AI coding product and suitable for a public listing.

Pier gives each coding-agent session its own cloud VM on the user’s AWS account, with parking, resuming, and repo setup built in.

A Go/Python tool for verifying AI audit logs, checking chain integrity, tamper detection, and whether a log is independently anchored outside the producing system.

Google Research article describing Science One Framework, an autonomous research prototype focused on verifiable AI-generated research through evidence chains and claim verification.

arXiv paper introducing ScientistOne, an autonomous research system built around a chain-of-evidence framework to keep claims traceable and audit research outputs for verifiability.

A Hermes plugin that uses adversarial LLM clones to confront sandbagging and reward-hacking behavior during agent sessions.

A CLI/library for line-level provenance in text edited by humans and agents, using version history to estimate which ranges are human-written versus machine-generated.

A research repo for reducing hallucinations in LLM-generated code using semantic triangulation, with setup, benchmarking, experimentation, and reproducibility instructions.

A web-based tester for MCP servers that inspects tools, resources, and prompts, measures latency, estimates token usage, and flags tool-poisoning risks.

OneRingAI is an open-source TypeScript library for building AI agents with integrations and graph memory. The repository shows a substantial codebase with docs, setup files, testing, MCP-related material, and references to supported models and multimodal capabilities.

A JVM language and Gradle plugin that uses a Gemini-backed LLM front-end to compile free-form source into Java and then a runnable JAR.

FAAAH is a dependency-free, OpenAI-compatible local proxy that routes requests through text files and AI coding agents. It is aimed at developers who want to reuse an existing agent subscription as a backend for tools that expect the OpenAI API.

Experimental page-aware Metal runtime and reproducibility harness for running oversized sparse mixture-of-experts models on Apple Silicon. The repo documents GPT-OSS 120B support, performance results on 64 GiB and 16 GiB machines, and the runtime/harness structure.

Claude Code documentation for cross-session messaging between independent sessions on the same machine and across machines via Remote Control. It explains when to use it, how messages are delivered, and how session permissions affect delivery.

Canonical Agent Notation (CANON) is an open-source controlled English for model prompting and agent-to-agent communication. The repository describes a small closed vocabulary, mandatory obligation markers, bounded loops, and a clear ask-rather-than-guess rule for making requests and agent messages less ambiguous.

Open-source Python project for execution-first multi-agent workflows built around YAML Charterfiles, runtime schema enforcement, and a boss-agent hierarchy. The README includes install commands, a short usage tour, and a small code example.

Rust-native desktop cockpit for coordinating multiple AI coding tools like Claude Code, Codex, and Gemini CLI in one interface. The page shows installation steps, supported platforms, core workflows, safety defaults, and release notes.

compactdiff snapshots, diffs, replays, and branches AI coding-agent sessions locally. It reads Claude Code and Codex logs natively, redacts sensitive data by default, and exports shareable replay HTML or forked JSONL branches.

MCP server that connects external AI agents to Revise for drafting, editing, reading, exporting, and sharing DOCX-compatible documents with tracked changes.

A technical case study on reducing token usage in a coding agent, including a model-free harness, wire-level measurement, and several concrete optimizations that cut a long run from 1.04M to 517k tokens.

A browser-based computer vision project that detects cursor hotspots in screen recordings using temporal motion cues, ONNX Runtime Web, and WebGPU. The repo includes a live demo, training scripts, synthetic data generation, and baseline comparisons against YOLO models.

GreatArrow.ai is an open beta workspace that gives Claude, ChatGPT, Cursor, Gemini, and other MCP-compatible clients shared memory plus action access across tools like Gmail, Slack, GitHub, Notion, and Jira.

A Git-based CMS for static site generators, packaged as a pnpm workspace with install, build, test, and demo instructions, plus documentation for usage and configuration.