egress-proxy

Open-source AI agent firewall for MCP security and agent egress. Scans mediated HTTP, MCP, A2A, and WebSocket traffic for exfiltration, SSRF, and prompt injection, and emits mediator-signed action receipts: verifiable audit evidence from outside the agent.

SourceryKit — counterspell for hallucinating agents. Python SDK that breaks the illusion on every tool call, API response, and MCP handoff before bad outputs propagate.

Egress-only, zero-dependency agent to plug private data sources into Weve.

Multi-tenant Kubernetes operator for self-hosted GitHub Actions runners. Scale-to-zero workers, per-tenant egress IP pools, and GPU priority scheduling across a shared ResourceQuota — an Actions Runner Controller (ARC) alternative.

一条命令在信创内网 GPU 盒上启动编码 Agent——零云上出口、零 SSH 隧道拼装、全量出口调用留痕可审。数据不出境的国企研发，第一次像用云端 runner 一样顺手。

Portable runtime policy and audit layer for AI agents - HTTP/HTTPS proxy enforcing egress policies, inspecting content, materializing secrets, and recording every decision.

Educational PoC demonstrating network isolation for Node.js apps using Docker internal networks and nginx as an egress proxy gateway.

Auditing egress proxy for AI agent deployments. Sits between agent and internet, intercepts HTTPS via mitmproxy, enforces an explicit allowlist, and logs every request to a tamper-resistant volume outside the agent. Comes with a built-in log viewer and live allowlist management.

Agent-specific network egress guardrail. Wraps untrusted LLM agent runtimes so they can't exfiltrate data or hold real secrets — the only route out is a default-deny proxy that swaps placeholder tokens for real secrets at the edge.