|
Xmore AI | AI Agent Security Assessment and Remediation
Test enterprise AI agents for over-broad permissions, unsafe tool use, prompt-injection exposure and data leakage, then implement and verify the fixes.
|
86 % |
|
Injentik — Offensive security for AI systems
We don't test your model. We test what your model can make your systems do. Independent AI red teaming rooted in offensive security, not ML.
|
86 % |
|
Security testing for AI agents that access data and take actions.
Security testing for AI agents that access data and take actions.
|
84 % |
|
TESTOS — AI Security Testing for LLM Apps, RAG, and Agents
Red-team LLM applications, RAG systems, and AI agents for prompt injection, jailbreaks, unsafe tool use, and data leakage before release.
|
84 % |
|
WardenAuth — Agent Security & Authorization Platform
Fine-grained authorization for developers, security teams, and platforms. RBAC + ReBAC + ABAC with flat-rate pricing. Agent Security for MCP. SDKs in TypeScr...
|
84 % |
|
NeverTrust.ai · Network-layer AI security
Network-layer AI security. An on-device agent intercepts AI traffic, scans response bodies with an embedded ML model, and enforces policy rules. No SDK. No c...
|
84 % |
|
Aegis — Adversarial security testing for AI models and agents
Aegis stress-tests your AI models and agents the way attackers will: multi-turn conversations, tool-call abuse, prompt injection. Outputs an audit-ready repo...
|
84 % |
|
WebMCP Audit: security audits for the agentic web
WebMCP lets any page register tools AI agents can call. We scan your tool registrations against the spec's threat model before agents ever touch them.
|
84 % |
|
Red Gambit · Security testing for teams building with AI
Security testing for AI products and the software around them, with a report your customers will accept.
|
83 % |
|
LLM Application and AI Agent Penetration Testing | Security Lab
Pentests for LLM apps and AI agents: prompt injection, RAG data leakage, MCP servers. Fixed price, reproducible report, retests for 60 days.
|
82 % |
|
AI Security Services for Startups and Enterprises | Mirawyn
Mirawyn secures your AI, the apps and code you build with AI, and your environment against AI-powered attackers. For solo founders through to enterprises.
|
82 % |
|
Red AI Team - AI Security Audits & Red Teaming
Automated security audits for AI chatbots, vibe-coded apps, and LLM-powered products. Prompt injection testing, vulnerability scanning, and remediation.
|
82 % |
|
KNOXLAB — Independent offensive security research
Independent offensive security research across web applications, APIs, and AI-enabled systems.
|
82 % |
|
runtimebreach — authorized application security testing
Authorized security testing for apps and platforms. Permission-first API and endpoint audits with evidence-backed findings.
|
82 % |
|
everbreach | Security testing for the age of AI attackers
AI is changing what attackers can find and exploit. EverBreach is developing security testing that follows attack paths across applications, APIs and connect...
|
82 % |
|
Elitor — Security testing for software built at AI speed
Independent, human-verified security assessments for apps built fast with AI coding tools. Fixed-price, standards-aligned, with a retest included — we find t...
|
82 % |
|
Blue Jay Application Security · Manual AppSec for SaaS & AI products
Manual application-security reviews for rapidly built SaaS and AI products. Auth, APIs, authorization/IDOR, tenant isolation, and AI-agent boundaries, with d...
|
82 % |
|
Agent Boundary Engagements | AI Red Teaming | Stonepath IT
Adversarial testing for AI agents, LLM apps and MCP integrations. Senior-led, fixed-scope, with evidence written for your customer's security review.
|
82 % |
|
breakyouragent: crash-test your AI agent
Paste your agent's endpoint. We throw hundreds of jailbreaks, prompt injections and tool-abuse attacks at it and show you exactly where it breaks.
|
82 % |
|
SteelCipher — Security Control Plane for AI Agents
SteelCipher helps companies safely deploy AI agents by controlling what agents can access, what they can do, what sensitive data they touch, and how that act...
|
81 % |
Detected social media networks