Safe runtime for autonomous on-chain AI agents: isolated sandboxes, Library skills, encrypted secrets.

Open-source agent sandbox infrastructure for parallel coding agents: isolated workspaces, MCP/CLI control, observability, and atomic publication.

Build system for AI coding-agent context. Generate deterministic, explainable context artifacts for Claude, Codex, Cursor, and MCP.

PixelFerrite is a C++20 security simulation framework with a Metasploit-style console for module workflows, image-based artifact generation, and workspace-isolated runtime for authorized research and validation.

Manufacturing dashboard where UI and LLM-generated reports adapt to the viewer role — engineers see evidence, managers see priorities, executives see decisions.

AI proposes. Humans decide. Source-available AI assurance and control plane for governed code change: agent identity, scoped authorization, policy gates, PR/CI evidence binding, replayable evidence, chained receipts, traceability, revision-bound assurance packages, independent verification, and human review.

A DeepSeek Harness plugin providing project workspace isolation: each project (task) maps to its own workspace and conversation, preventing rules, memory, and context from leaking across projects — so multiple projects can progress in parallel.

Per-session workspace scoping for DeepSeek Harness — expose only selected directories to agents with focused and isolated enforcement.

OpenClaw plugin that prevents social agents from reading the main agent's workspace, config, and credentials. Fills the gap where workspaceOnly enforces writes but not reads. Eight audit versions, 6 blockers fixed.

OpenClaw plugin for durable specialist AI agents with isolated workspaces, memory, state, and /persona routing.

Codex 会话工作区隔离本地插件