DOS Kernel - trust substrate for agent fleets (stdio via PyPI)
@anthony-chaudhary
DOS Kernel - trust substrate for agent fleets (stdio via PyPI) について
Catch your AI agents when they lie about what they shipped — verifies claims against git instead of believing the agent.
基本情報
設定
ツール
ツールは検出されませんでした
ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。
概要
What is DOS Kernel - trust substrate for agent fleets (stdio via PyPI)?
DOS Kernel is a trust substrate for agent fleets that verifies agents' claims of shipped work against real-world artifacts—starting with git history. It runs as a referee via the dos CLI, installed as a single Python package (dos-kernel). It is built for teams running AI agents on code repos who need an objective verdict on whether work actually landed.
How to use DOS Kernel - trust substrate for agent fleets (stdio via PyPI)?
Install with pip install dos-kernel, then run dos init --hooks auto to wire into your existing agent runtime (e.g., Claude Code, Cursor) or dos quickstart for a throwaway demo. Use dos verify <plan> <phase> to check a claim against git commits: SHIPPED (exit 0) or NOT_SHIPPED (exit 1). The verdict can gate further actions in CI, MCP hosts, or fleet frameworks.
Key features of DOS Kernel - trust substrate for agent fleets (stdio via PyPI)?
- Verifies agent claims against git history, never the agent's narration.
- Detects silent collisions between agents racing on the same files.
- Flags spinning runs and saves compute by quitting doomed tasks.
- Works on any plain git repo with zero configuration.
- One runtime dependency: PyYAML (MIT license).
- Integrates via MCP, runtime hooks, CI, and fleet frameworks.
Use cases of DOS Kernel - trust substrate for agent fleets (stdio via PyPI)?
- Catching agents that falsely report shipping features (e.g., "login endpoint shipped" when no commit exists).
- Preventing silent file overwrites and collisions in multi-agent fleets.
- Saving fleet compute by automatically stopping runs that produce no landable work.
- Improving acceptance precision by rejecting unbacked claims before they pollute code.
FAQ from DOS
「AI とエージェント」の他のコンテンツ
Shell and Coding agent for Claude and other mcp clients
rusiaamanShell and coding agent on mcp clients
meGPT - upload an author's content into an LLM
adriancoCode to process many kinds of content by an author into an MCP server
MCP Server - Remote MacOs Use
baryhuangThe only general AI agent that does NOT requires extra API key, giving you full control on your local and remote MacOs from Claude Desktop App
1Panel
1Panel-dev🔥 1Panel is a modern, open-source VPS control panel — and the only one with native AI agent support. Run Ollama models, deploy OpenClaw agents, and manage your entire server stack from one clean web interface.
🔎 GPT Researcher
assafelovicAn autonomous agent that conducts deep research on any data using any LLM providers
コメント