truefoundry/trueforge

Source
GitHub
First trending
Category
Agents
GitHub stars
5,713
Main language
TypeScript
Website
trueforge.dev (opens in a new tab)

This page introduces an external open-source repository. It is not an HDATF product.

truefoundry/trueforge

What it does

TrueForge is an open-source agent harness that handles model calls, MCP tools, skills, sandboxing, approvals, context management and session state. It offers a chat UI, an HTTP API with a TypeScript SDK, and an embeddable UI SDK.

How it helps ATF

Worth comparing directly with Harness, since it also runs an agent loop and keeps session state. Its tool approval checkpoints, sandbox handling and context compaction could be a reference.

License

MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.

More in this category

  • yjh051108/dsh-routing-suite
    A plugin suite for DSH: a runtime injector with hot reload, a task-aware reasoning-mode router, and a two-level task protocol with planned steps, check-ins, a red-team gate and final verification.
  • chaitanyagiri/munder-difflin
    A local desktop multi-agent harness that wraps terminal coding CLIs such as Claude Code, Codex and Gemini CLI. Each agent gets long-term memory and a mailbox, and a coordinating agent routes work between them on a 2D office view.
  • openJiuwen-ai/jiuwenswarm
    A multi-agent system where a leader agent splits complex tasks across teammate agents, including scripted multi-stage workflows with human checkpoints. Skills are revised automatically from error signals, and it is reachable from a browser, terminal or chat apps.
  • walkinglabs/learn-harness-engineering
    A project-based course on harness engineering, covering the environment, state management, verification and control that make AI coding agents work reliably. It also breaks down the harness designs of Pi, Claude Code, Codex and DeepSeek.
  • sapientinc/PRAXIST
    An autonomous research system for measurable, computer-executable research. It coordinates parallel research peers, lets each task own its evaluation, keeps durable evidence and carries synthesis from one generation of work to the next.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.

View on GitHub (opens in a new tab)