Tools
Agentic AI risk checker — planned
Status: planned. This tool is not available yet. Join the waitlist and we will let you know when an early version is ready to test. We do not have a release date to share.
What it is intended to do
Agents carry risks that model-centric assessments miss: what they can touch, what they can do without asking, and how quickly they can be stopped. The agentic risk checker is planned to review a single agent across:
| Area | Questions it will ask | Controls |
|---|---|---|
| Tool inventory | Which tools, scopes and data stores can the agent reach? | EI-07 |
| Least privilege | Is access default-deny and limited to the declared purpose? | IS-11 |
| High-impact actions | Which operations are irreversible, external or financial, and do they require approval? | IS-13 |
| Injection exposure | Does the agent read untrusted content that could carry instructions? Has it been tested? | IS-14, IS-01 |
| Containment | Can the agent, a tool or a tenant be disabled quickly? Has that been tested? | IS-15 |
| Limits | Are there rate, token, tool-call and cost limits? | IS-16 |
| Logging | Are tool calls and policy decisions logged tamper-evidently? | SE-08 |
VERITRAIL controls cited
- EI-07 Agent and tool registryNIST AI RMF: GOVERN 1.6, MAP 4.1ISO/IEC 42001: A.4.4, A.6.2.3EU AI Act: Art. 11
- IS-11 Least-privilege tool accessNIST AI RMF: MEASURE 2.7, MANAGE 2.2ISO/IEC 42001: A.6.2.6, A.4.4EU AI Act: Art. 15, Art. 14
- IS-13 Human approval for high-impact actionsNIST AI RMF: GOVERN 3.2, MANAGE 2.2ISO/IEC 42001: A.9.2, A.6.2.6EU AI Act: Art. 14
- IS-14 Prompt-injection and data-leakage testingNIST AI RMF: MEASURE 2.7ISO/IEC 42001: A.6.2.4EU AI Act: Art. 15
- IS-15 Kill switch and rollbackNIST AI RMF: MANAGE 2.4ISO/IEC 42001: A.6.2.6EU AI Act: Art. 14
Framework mappings are indicative VERITRAIL interpretations (Control Library v1.0.0); they are not legal advice and do not establish conformity or compliance.
Available today
The guardrail generator already drafts a tool permission matrix and runtime policy for an agent, and the maturity self-assessment flags untested agents as a critical gap.
Join the waitlist
Last reviewed: