Agentic Highway is the trust and safety layer for the agentic world. Scan your machine with Proov, our free CLI, and see every agent, skill, MCP server, and prompt — with trust scores and risk findings — before anything ships.
Get notified when we launch publicly:
web-search-agent
by Anthropic
0+
checks behind every review
0
review outcomes teams can act on
0
steps from submission to search
0
directory to check before install
The Product
Search analyzed agents, tools, and integrations to see what’s been checked before your team installs, approves, or ships it.
Check what you’re bringing into your stack before you install it. Search analyzed agents, tools, and integrations by score, status, and publisher.
web-search-agent
Anthropic
file-manager-pro
DevToolsCo
code-review-assistant
SecureAI Labs
data-pipeline-builder
CloudNative
email-automation-kit
MailFlow
db-query-optimizer
SQLWorks
4 of 6 listings verified
View full directory →The Process
Step 1
Submit
A publisher submits a GitHub repo or package so Vettd can analyze what teams would actually install.
Skill Publisher
Step 2
Scan
Vettd checks repo protections, CI/CD setup, and dependency hygiene, then reviews package structure, permissions, and unsafe patterns.
Vettd Scanner
Step 3
Score
Findings roll into a score and status badge (Vettd, Conditional, or Flagged), with clear results publishers can act on.
Vettd Platform
Step 4
Discover
Teams browse the directory to compare analyzed agents, tools, and integrations before they install, approve, or deploy.
Developers
Why trust the badge? It's based on repeatable automated checks, not marketing copy or human opinion. Vettd evaluates repo protections, dependency hygiene, package structure, and unsafe patterns, then shows the findings behind each result. We're expanding coverage over time, including framework mapping and continuous monitoring.
For Enterprises
AI agents are landing in enterprise stacks faster than most review processes can keep up. We pair the Proov platform with architecture-led consulting so you can adopt AI agents with confidence, not guesswork.
Enterprise-grade Proov scanning with continuous monitoring
Architecture-led consulting with a 5-phase engagement model
Framework mapping and remediation guidance for regulated teams
See the full enterprise platform and consulting engagement details.
For Developers
Submit a GitHub repo or upload a package to Vettd and get an immediate trust review across ~50 checks. See what passes, what needs work, and whether you qualify for a trust badge without waiting on a manual review. Results arrive in seconds, so you can fix gaps while the context is still fresh.
The Vision
Two products, one mission: make it easier to adopt AI tools with confidence and harder for risky agents or packages to slip into production.
The Trust Directory
A public directory for AI agents, tools, and integrations. Teams can submit repos or packages for analysis, then share a searchable listing, trust score, and portable badge wherever credibility matters.
The Analysis Engine
The analysis engine behind Vettd. Every submission runs through ~50 checks across code quality, dependency risk, permission scope, and behavior signals — producing a trust score and findings teams can actually review.
Workflow Orchestration
A CLI for orchestrating multi-step AI workflows with a plugin ecosystem and first-party integrations for OpenAI, Anthropic, and more. Kelvin focuses on execution; Vettd is the trust and verification layer around the ecosystem.