OpenClaw Security Guide
Back to Threat Intel
sourcepostAgent: OpenClaw

Show HN: Probus, AI vuln scanner

A May 5, 2026 Hacker News post announcing Probus, an AI-assisted vulnerability scanner. The author says the tool was run against projects they use and lists reported findings in n8n, Vercel AI SDK, LangGraph.js, browser-use, and Haystack.

agentic-aicoding-agentsvuln-scanninginput-validationschema-validation

Date

May 5, 2026

First Seen

May 5, 2026

Last Reviewed

May 8, 2026

Publisher

Hacker News

Source Type

post

View source

Related reading

OpenClaw Security Guide

A practical baseline for local binding, scoped credentials, sandboxing, runtime checks, and Armorer Guard.

Securing OpenClaw with Armorer Guard

How Armorer wraps OpenClaw with managed setup, Docker hardening, health checks, approvals, and Guard-backed scanning.

Get email updates

Get reviewed Armorer threat-intel updates when new findings are published.

Source Summary

What It Contains

A May 5, 2026 Hacker News post announcing Probus, an AI-assisted vulnerability scanner. The author says the tool was run against projects they use and lists reported findings in n8n, Vercel AI SDK, LangGraph.js, browser-use, and Haystack.

Extracted Claims

  • The author reports three Vercel AI SDK findings, including a runtime schema bypass in ToolLoopAgent.
  • The post links the Vercel AI SDK findings to GitHub pull requests, including PR #14750 for callOptionsSchema enforcement.
  • The described scanner design separates analyst, researcher, and QA agents, with each agent running in its own query session and a filesystem sandbox scoped to the target repository.

Evidence Quality

Useful fresh public discussion and discovery context, but it is not the primary vulnerability record. The linked upstream GitHub pull request is the stronger evidence for the Vercel AI SDK issue.

Follow-Up

  • Prefer upstream project PRs, advisories, or releases for authoritative remediation details.
  • Track whether the scanner's multi-agent verification pattern becomes a reusable defensive control or benchmark for agent-framework security testing.