01 / Assessment service
Aspherity
Adversarial assessment of agents with access to data and connected actions.
THE DIRECTORY / ASSESSMENT SERVICESCompare AI security
Compare AI security
assessment providers.
A researched starting point for your shortlist. Each profile separates public provider statements from the questions you still need to ask.
Listed alphabetically. Unknown capability does not mean unavailable.
02 / Hybrid service
Astra
Human-led AI application penetration testing paired with a testing platform.
03 / Assessment service
Bastet Security
An MCP architecture assessment covering access controls and integration trust.
04 / Assessment service
Bishop Fox
AI security assessments spanning agents, retrieval pipelines and surrounding application infrastructure.
05 / Hybrid service
Casco
AI application and agent workflow testing, with an optional security-engineer review.
06 / Hybrid service
Cobalt
Human-led AI application pentesting with agent workflow and retrieval-layer coverage through the Cobalt platform.
07 / Assessment service
DeployOpen
Fixed-scope agent testing in a controlled environment with written rules of engagement.
08 / Hybrid service
NetSPI
Expert-led AI and LLM testing with custom evaluations and continuous testing options.
09 / Assessment service
Optiv
MCP security assessments of server deployments and their enterprise integrations.
10 / Assessment service
OSec
Senior-led AI red teaming covering tool-enabled agents, MCP servers, retrieval pipelines and agent-to-agent interactions.
11 / Assessment service
Themis
Agent assessments built around a capability map and an owner-confirmed authority baseline.
12 / Assessment service
Trail of Bits
AI/ML assessments covering deployed agents, their pipelines and exploitable trust boundaries.
No matching providers
Try a broader search or reset the filters. Missing public documentation is not proof a provider lacks a capability.