Agent security, evaluation and observability
Explore security operations, agent inspection, tracing and evaluation.
8 records: 4 organization · 3 product · 1 tool
Source-linked research and editorial guidance. Capabilities are not independently tested unless a profile explicitly includes a reproducible test record. Sources checked November 13, 2026.
What to compare
Separate observations from verdicts. Compare the scope of each scan or evaluation, tested versions and how failures are reproduced.
Records in this category
Alphabetical; not a ranking. Each card shows the record's supplied type and profile depth.
7AI
Agents investigate alerts, hunt threats and support customer-controlled response.Arize Phoenix
Open-source, locally runnable or self-hosted observability and evaluation for agents.Braintrust
Agent evaluation and observability platform with task and trajectory testing.Dropzone AI
SOC agents investigate alerts and threat leads while analysts retain verdict ownership.Langfuse
Open-source tracing, evaluation and usage/cost observation for AI applications and agents.LangSmith
Platform for evaluating agent behavior and collecting feedback before and after deployment.Snyk Agent Scan
Scanner for agent components, MCP servers and skills.Torq
Agentic security operations spanning triage, investigation and response.Other categories: Coding agents · Browser and computer agents · Customer support agents · Sales and customer research agents · Business operations agents · Finance and accounting agents · Legal and professional agents · Healthcare workflow agents · Procurement agents · Science and research agents · Agent frameworks and platforms · MCP servers and integrations · Agent skills · Agent payments and commerce · Agent discovery and task markets