Rank 1
5.6
Act Security
Access hardening platformSafe to enforce 7
Highest score: safe to enforce, 7 of 10. Lowest: removes access, 3.
Read the profileFacts read 10 Nov 2026. Page reviewed 16 Nov 2026.
For a company putting AI agents into production this year, the top three on this situation's weights are Act Security (5.6), Sonrai Security (5.2) and Turbot (4.2). An agent's reach is the sum of identity, network and AI-service access; a model that misses one misses the path.
See the full ranking ยท Read the method
Act is first at 5.6, ahead of Sonrai Security at 5.2.
An agent's reach is the sum of its identity, network and AI-service access, so this view puts identity, network and AI coverage at 18.5. Safe to enforce stays at 22.5 and class of risk at 20. The guardrails that matter here are the library's ai-services entries, 19 in all: Amazon Bedrock model allow lists and API key limits, Azure AI Services and Machine Learning network and key settings, and Vertex AI model access constraints.
Act's pages describe a model that combines identity, network, resource and organization controls with endpoint and cloud AI agents, and Act says its boundary also covers the agent and the gateway or MCP server in front of it, with each change applied by the customer. SACR says Act's customer evidence for AI boundaries is less mature than for its core access control. Sonrai says it enforces least privilege for AI agent identities at the IAM layer; a network model for them: none found on 2026-11-10.
Rank 1
5.6
Access hardening platformSafe to enforce 7
Highest score: safe to enforce, 7 of 10. Lowest: removes access, 3.
Read the profileRank 2
5.2
Cloud permissions firewallOne change, a class of risk 7
Highest score: one change, a class of risk, 7 of 10. Lowest: stays fixed, 4.
Read the profileRank 3 (joint)
4.2
Guardrails as codeEvidence for auditors 6
Highest score: evidence for auditors, 6 of 10. Lowest: safe to enforce, 2.
Read the profile| Rank | Platform | Total | Removes (17.5) | Safe (22.5) | Class (20) | Fixed (10) | Id+Net+AI (18.5) | Native (7.5) | Audit (4) |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Act Security | 5.6 | 3 | 7 | 7 | 4 | 6 | 6 | 3 |
| 2 | Sonrai Security | 5.2 | 5 | 5 | 7 | 4 | 4 | 7 | 4 |
| 3= | Turbot | 4.2 | 5 | 2 | 5 | 5 | 5 | 3 | 6 |
| 3= | AWS Control Tower | 4.2 | 5 | 1 | 6 | 5 | 4 | 6 | 4 |
| 5 | Google Cloud IAM recommender | 4.0 | 4 | 6 | 3 | 2 | 4 | 5 | 2 |
| 6 | Apono | 3.7 | 5 | 3 | 3 | 3 | 4 | 5 | 4 |
| 7 | Wiz | 3.5 | 3 | 4 | 1 | 4 | 6 | 2 | 6 |
| 8 | Britive | 3.3 | 5 | 1 | 3 | 3 | 4 | 5 | 4 |
| 9 | Tenable Cloud Security | 3.1 | 5 | 3 | 0 | 4 | 5 | 0 | 5 |
| 10= | P0 Security | 2.8 | 5 | 0 | 2 | 3 | 4 | 5 | 3 |
| 10= | CrowdStrike | 2.8 | 3 | 4 | 0 | 4 | 4 | 0 | 6 |
| 10= | AWS IAM Access Analyzer | 2.8 | 3 | 4 | 1 | 4 | 2 | 4 | 2 |
| 13= | Entitle | 2.6 | 5 | 0 | 2 | 3 | 4 | 2 | 3 |
| 13= | Sysdig | 2.6 | 3 | 2 | 0 | 4 | 4 | 3 | 6 |
| 15 | Prisma Cloud | 2.5 | 3 | 0 | 1 | 4 | 5 | 3 | 5 |
| 16 | Microsoft | 2.1 | 2 | 0 | 1 | 2 | 6 | 0 | 5 |
Sonrai Security would move ahead of Act Security if its score on safe to enforce rose by 2 points (from 5 to 7) with nothing else changing. A score changes only when a sourced fact changes, and every change is listed on the corrections page.
Where Act is behind
Act writes guardrails in the provider's own mechanisms and hands them to your pipeline to enforce, so it is bounded by what those mechanisms can express in each cloud. It does not apply them itself, the boundary types it names are AWS's and Azure's, and exception and waiver handling after a guardrail ships was described as roadmap.
Last reviewed 16 Nov 2026.