About · v3.3.0 · Aug 8, 2026

PipeLab is the company behind Pipelock.

Pipelock is an open-source agent firewall for MCP and agent traffic. It sits between AI agents and the internet and blocks secret leaks, unsafe tool traffic, and prompt-injection responses.

Also published: Gauntlet Proof Comparisons Technical guides

31,000+

Test cases

47

Releases

214

Bench cases

11

Scanner layers

Go

Single binary

Corpus snapshot verified against agent-egress-bench@fb75dbf6e381.

Origin

From a homelab problem to a public security tool

PipeLab started as a homelab project, not a startup. It grew because a real operational problem needed a real fix.

  1. 2024 · Homelab

    A plumbing business wanted to automate operations. The founder spun up a homelab and started building AI agents to take dispatch, scheduling, and back-office work off the calendar.

  2. The problem

    Those agents had shell access, API keys, and unrestricted network. One mis-prompted call, one poisoned tool response, and credentials walk out the door. Nothing on the market sat between the agent and the internet.

  3. The fix

    Pipelock was built to fill the gap. A single Go binary, Apache 2.0, scanning mediated HTTP, WebSocket, and MCP messages for credential leaks, prompt injection, SSRF, and tool poisoning.

  4. Today

    Pipelock is listed in the CNCF Landscape under Security & Compliance, and publishes coverage mappings for OWASP MCP Top 10, OWASP Agentic Top 10, MITRE ATLAS, EU AI Act, NIST AI RMF, HIPAA, and SOC 2.

Ship

What PipeLab ships

A small ecosystem around AI-agent runtime security. All open source where it can be, all evidence-backed where it counts.

Pipelock

The open-source agent firewall. A single Go binary that sits between an AI agent and the internet, scanning mediated messages through a 11-layer pipeline.

LicenseApache 2.0
Tests31,000+
Releases47

Agent Egress Bench

A tool-neutral attack corpus for validating any agent egress proxy. Public methodology, public attack cases, public results.

Cases214
Categories18
LicenseApache 2.0

Corpus snapshot verified against agent-egress-bench@fb75dbf6e381.

Pipelock Rules

Community detection patterns shipped as signed YAML bundles. Hot-reloadable. Adds DLP, prompt-injection, and tool-poisoning coverage without a redeploy.

DLP patterns65
SigningEd25519
ReloadHot

CNCF Landscape

Pipelock is listed in the Cloud Native Computing Foundation's Landscape under Security & Compliance. Coverage mappings for OWASP MCP Top 10, OWASP Agentic Top 10, MITRE ATLAS, EU AI Act, and SOC 2 are published as public guides.

CategorySecurity & Compliance
Frameworks7
FormatPublic guide

Philosophy

How PipeLab works

01 / Build it yourself

Build it yourself. Understand it yourself.

If it breaks, you know why. If it works, you know how. Every line of the scanner pipeline is auditable and tested in the open.

02 / Honest engineering

Open-source core. Honest pricing. Real engineering.

The binary is Apache 2.0. The bench is public. The rules repo is signed and versioned. Pro features pay for multi-agent coordination, not for access to what should be free.

No cloud dependencies No VC funding No growth hacking