Illustrative sample. This is not the output of a real scan — it exists to show the shape of a Sentinel report.

Enterprise AI Trust Report

Sample scan result

orders-agent · acme-corp/checkout-agent@a9f3e21

Signed & independently verifiable
81/ 100Overall trust
88/ 100Security
74/ 100Trust
92/ 100Quantum
26RISKRisk

Trust radar

IdentityAuthorityDelegationPolicyAuditEvidenceQuantum

Trust trend · 30 days

+13

Trust improved after least-privilege recommendations were applied to two over-scoped tools.

Trust dimensions

5/7 pass

Identity

Signed identity present and verifiable

PASS

Authority

Two tools granted broader scope than used

WARNING

Delegation

Delegation bounded, no re-delegation

PASS

Policy

No enforced guardrail on payment actions

FAIL

Audit

Actions recorded in tamper-evident order

PASS

Evidence

Each action independently provable

PASS

Quantum Ready

Hybrid PQ key exchange, ML-DSA-65 signing

PASS

Findings & recommended fixes

CriticalAgent Trust

Payment tool has no human-approval gate

Fix — Insert an approval chain for value-transfer actions
HighSecurity

API key with write scope in environment

Fix — Rotate to a read-scoped, short-lived credential
MediumAgent Trust

Filesystem tool granted unbounded path

Fix — Restrict to the working directory (least privilege)
LowQuantum

Dependency pins a pre-PQC TLS library

Fix — Upgrade to the hybrid PQ-capable release

Compliance mapping

OWASP LLM
NIST AI RMF
ISO 42001
ISO 27001
SOC 2
PCI DSS
HIPAA
GDPR

Run this on your own AI.

Point Sentinel at a repository, a container, or an MCP server and get your own report.

Scan My AI