Assay is an MCP server tool by Rul1an designed for Continuous Integration (CI) environments. It connects to agent systems, providing CI-native evidence for observed agent behavior. This free mcp category tool helps enforce MCP policies, generate evidence receipts, and support Trust Basis claims. Assay also aims to enforce egress directly within the Linux kernel using eBPF/LSM, acting as a crucial bridge between AI applications and external capabilities. It helps deny risky tool calls before execution and proves what ran with verifiable evidence.

What does Assay connect to?

Assay functions as an MCP server tool that establishes connections with agent systems. It serves as a bridge between AI applications, which act as clients or hosts, and various external capabilities. Communication primarily follows the JSON-RPC 2.0 protocol. It supports several transport protocols, including standard input/output (stdio), streamable HTTP, and Server-Sent Events (SSE). While designed for compatibility with general AI clients like Claude Desktop and Claude Code, specific AI clients confirmed compatible with Assay are not explicitly stated.

Core features and tools

Assay provides several core functionalities for agent systems. It enables MCP policy enforcement, allowing risky tool calls to be denied before they run. The tool generates verifiable evidence, evidence receipts, and Trust Basis claims to prove agent actions. Furthermore, Assay is designed to enforce egress in the kernel using eBPF/LSM on Linux systems. This capability helps ensure observed agent behavior adheres to defined policies, providing crucial artifacts for review and accountability within CI workflows.

Who is it for?

Assay targets developers, engineers, and AI consultants who work with agent systems, especially within Continuous Integration contexts. Its design supports those who need to enforce strict policies, generate auditable evidence of agent actions, and manage kernel-level egress. Professionals focused on security, compliance, and verifiable execution in AI-driven environments will find Assay particularly useful. Skip this if your work does not involve agent systems or CI policy enforcement.

How to get started

No credentials are required. Assay supports communication via standard input/output (stdio) and streamable HTTP.

Use cases

• Enforcing MCP policies by denying risky tool calls before execution. This ensures agent actions comply with security and operational guidelines in Continuous Integration pipelines, preventing unauthorized operations.
• Generating verifiable evidence, evidence receipts, and Trust Basis claims to prove agent actions. These artifacts are essential for audit trails, compliance, and establishing trust in automated processes.
• Enforcing egress policies directly within the Linux kernel using eBPF/LSM. This provides fine-grained control over agent system network traffic, ensuring communication only with approved external destinations, enhancing security.

Assay by Rul1an is a valuable MCP server tool for CI environments, focused on policy enforcement and verifiable evidence for agent systems. It helps manage agent behavior and kernel egress, providing essential artifacts for review. While specific integration details require further exploration, its core capabilities are clear.