GitHub listed eleven projects this Friday morning, every one of them an AI tool, agent add-on or prompt collection. Six had been featured in the past week and none showed a change big enough to bring it back, so all four picks are first-time entries. CopilotKit's OpenBot leads as a self-hosted template for governed company agents, OpenMausBot turns local Claude, Codex and Grok tools into a roster of chat bots with their own computers, claude-mem gives coding agents memory across sessions, and Claude Code Game Studios adds a whole studio's process to game projects. The remaining new entry, a collection of model jailbreak prompts, was left out again. Curation order reflects practical interest, not GitHub rank. Trending is not a release date or an endorsement.
Picks are listed in curation order, not GitHub’s rank; each card shows the rank GitHub displayed when the page was checked. Stars measure attention, not security, quality or fitness for use, and no project here was installed or run. “Stars today” is GitHub’s own displayed figure for its daily window, not an audited rolling 24-hour count.
Pick 1 · AI agents / self-hosted platformGitHub rank #9 of 11
What it does. A self-hosted template for company AI coworkers from the CopilotKit team. Each bot gets its own browser, files and only the tools an administrator grants; Docker Compose runs the app, API, routine worker and PostgreSQL, and any agent that speaks the open AG-UI protocol (LangGraph, Mastra, CrewAI, Pydantic AI, Google ADK or hand-written) can be plugged in. It ships with a General Assistant and a Knowledge bot, plus example fintech bots.
Caveat. The README calls it beta, and a template rather than a product. It requires a CopilotKit Intelligence project and license (a free plan exists; the local self-hosted evaluation licence lasts 30 days and is a macOS preview), plus your own model key. The example settings run in single-user mode, which admits every request as one administrator, so turn on sign-in before anyone else can reach it. The governance claims are the maintainers'; we did not run it.
Attention, not endorsement
GitHub showed
“247 stars today” as displayed for GitHub’s daily window; not an audited rolling 24-hour count
What it does. A desktop chat app (macOS, Windows, Ubuntu 24.04) where each AI bot is a contact with its own model and computer. The bots run on the Claude, Codex or Grok command-line tools already signed in on your machine, through a small server on 127.0.0.1 that keeps transcripts and keys in ~/.openmausbot. A bot can use a cloud Linux desktop, a local VM or, after explicit opt-in, your own computer, plus connected apps through Composio.
Caveat. Early software: Ubuntu builds are beta and calls are macOS-only. The cloud computers come from Boat, a paid service after its trial, and connected apps need Composio, each with its own account and terms. Bots can create schedules and webhooks without a confirmation step (the chat shows an Undo card), and anonymous usage analytics are on until you switch them off. Letting agents drive a real computer is a security decision; we did not install it.
Attention, not endorsement
GitHub showed
“1,199 stars today” as displayed for GitHub’s daily window; not an audited rolling 24-hour count
What it does. A persistent memory plugin for coding agents, built first for Claude Code and now also installable for OpenCode, Codex through T3 Code, Pi and others. Lifecycle hooks record what the agent does in a session, a model compresses that into summaries stored in local SQLite with Chroma vector search, and later sessions get relevant context back through a search skill and four MCP tools.
Caveat. It works by capturing everything your agent does, so secrets in tool output can end up in its database unless you wrap them in its "private" tags. Some model has to summarize that activity: the installer suggests signing in to a hosted claude-mem observer (30-day free trial, after which it falls back to your Anthropic plan unless you subscribe), or you can choose your own OpenRouter, Gemini or Anthropic access. Cloud sync to cmem.ai is optional. It installs hooks into your agent; we did not run it.
Attention, not endorsement
GitHub showed
“336 stars today” as displayed for GitHub’s daily window; not an audited rolling 24-hour count
What it does. A project template that turns Claude Code into a game studio of 49 specialized agents: directors, department leads and specialists for design, programming, art, audio, narrative, QA and release, with engine-specific sets for Godot 4, Unity and Unreal Engine 5. The README lists 74 slash-command skills, 15 hooks that validate commits, pushes and assets, 14 path-scoped coding rules and 62 document templates.
Caveat. Built for Claude Code only. The three director agents ask for Opus and others for Sonnet, which can raise usage costs on a metered plan. The 15 hooks are Bash scripts that run on your commits and pushes and can block some of them, so read them before adopting. The rigor comparison is the maintainer's: it says four games were built but the table shows two. It writes no game by itself; we did not install it.
Attention, not endorsement
GitHub showed
“353 stars today” as displayed for GitHub’s daily window; not an audited rolling 24-hour count
Every repository GitHub listed, in GitHub’s order, with the figures it displayed when the page was checked (): 4 picked today, 6 covered in an earlier digest, 1 not covered. For projects covered before, growth compares the total stars at our first pick with today’s displayed total.
TypeScript · 709 stars today · 4,035 stars in total
A local-first desktop workspace for persistent AI teammates. Run Codex, Claude, and Grok with dedicated workspaces, task queues, file sharing, browser control, and agent-to-agent collaboration.
Covered before — first picked 27 Oct 2026 at 1,697 stars; now 4,035: +2,338 stars (+137.8%) in 3 days, about 779 a day.
TypeScript · 336 stars today · 105,064 stars in total
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
TypeScript · 247 stars today · 8,328 stars in total
Open-source AI coworkers that each get a computer of their own: a browser, files and tools, with every action decided before it happens and recorded after. Bring any AG-UI agent.
Every project picked in an earlier digest, biggest growth first. “Then” is the total at our first pick’s metadata check; “now” is GitHub’s API star count checked . Growth measures attention, not quality or safety.
All 11 listing entries were inspected. Exact primary public GitHub repository metadata and recent releases were checked for all 11 at about 07:01 UTC, and README files for the four picks; all were active and unarchived. Ten had an API-detected license (nightly-labs/openbot as NOASSERTION); togg53192-cmd/jailbreaks had none.
Six entries featured in the previous seven days were left out without a material change: HKUDS/DeepTutor, obra/superpowers and elayadesign/ai-design-skills (28 October), nightly-labs/openbot (27 October), JetBrains/thinkrail (26 October) and pbakaus/impeccable (25 October). Since those cards they show patch and feature releases or no new release.
JetBrains/thinkrail released v0.3.0 on 29 October, after its 26 October card. Its notes list interface translation, a merged Files panel, GitHub review details in the top bar and a change titled "unconditional PII-free PostHog error tracking". We note it here rather than repeat the card; HKUDS/DeepTutor v1.7.0 (28 October) was noted yesterday.
togg53192-cmd/jailbreaks, a collection of prompts written to override model safety policies, was again not picked: GitHub detects no license, and its purpose is to defeat safeguards rather than to serve as a practical tool.
Today's listing had no security, infrastructure or non-AI projects, so the picks are less varied than usual; we did not reach outside the daily listing to fill that gap.
The preceding seven calendar dates, 23-29 October, were checked for local history; all seven editions exist.
The official listing was captured at 08:00:34 Europe/Paris. Literal stars-today labels are GitHub's displayed period figures; API total stars and README evidence were checked separately at the times shown per card. They are not an independently audited rolling 24-hour growth measurement.
Feature, safety, cost and performance claims are the maintainers' own, including claude-mem's token-savings estimate, OpenBot's action gateway and Claude Code Game Studios' rigor comparison. No candidate code, installer, plugin, hook, skill or model was installed, downloaded or run, and none of this is a security audit or endorsement.