Search

Sentinel capability sandbox: generate an OS sandbox profile from approved capabilities and run lifecycle scripts under it (macOS Seatbelt / Linux bubblewrap, deny-by-default).

Sentinel namespace-claim steward: DNS TXT claim verification, renewal/freeze lifecycle, timelocked transfers, and Ed25519-signed claim/retraction corpus releases.

Sentinel CLI: pre-install audit verdicts, whole-tree lockfile audits, registry-redirected npm/npx, policy tooling, and sandbox-enforced installs.

Sentinel CI runner: self-booting dependency-tree audit for GitHub Actions (sentinel-ci) — audits a lockfile, writes a CycloneDX SBOM, and posts a PR verdict.

Developer CLI for Lovrabet apps, datasets, pages, SQL, BFF, database connections, skills, and AI agent workflows.

Local-first AI code agent in your terminal. Inspired by Claude Code and Codex CLI.

General purpose 'deep agent' with sub-agent spawning, todo list capabilities, and mock file system. Built on LangGraph.

Pydantic AI capability for chDB, the in-process ClickHouse engine: analytical SQL tools with engine-level read-only safety and typed engine errors mapped to ModelRetry.

LLM model: Kwaipilot: KAT-Coder-Air V2.5

LLM model: Kwaipilot: KAT-Coder-Pro V2.5

Landing page feedback plus an auto-generated improved version

4-bit LoRA with Unsloth, small and readable

Fine-tune in 30 lines, free on Colab

Starter agent, structured outputs, tools (built-in, function, third-party, MCP), memory, callbacks, plugins, and multi-agent patterns. Model-agnostic.

Starter agent, function calling, structured outputs, tools, memory, evaluation, handoffs, swarm orchestration, and routing logic.

Local daemon that bridges Claude Code hooks to AgentMem.

A powerful and flexible framework for building, orchestrating, and deploying multi-agent systems.

This package will be released in the near future. Stay tuned!

Embedded long-term memory sessions for the OpenAI Agents SDK

AI model: mradermacher/minicpm5-1B-GLM-5.2-Agentic-v9-GGUF