Trust Center: SOC 2 Type I

Trust Center

See what your agents did, and what Darkhunt stopped

Darkhunt records every step an agent takes and what it decided about each one, so you can answer anyone who asks.

One claim, four decisions

A claims agent asked to send one lab result, as it appears in the record.

Open a step to see what it touched.

One insurance claim

Claims agent

2 allowed1 flagged1 stopped

01Opened the patient’s claimAllowedPart of the task

Data touched: the claim Tool used: patient record lookup

02Read the lab resultAllowedThe item the insurer asked for

Data touched: one lab result Tool used: patient record lookup

03Read medications and clinical notesFlaggedPermitted, but more than this task needs

Data touched: medications, clinical notes Tool used: patient record lookup

04Tried to send the full history to the insurerStoppedMore than the task asked for. Sent to a person to review.

Data touched: diagnoses, medications, clinical notes Tool used: insurer upload Nothing was sent. The lab result can go once a person approves.

What each entry holds

Every prompt, response, tool call and data access is logged, with what Darkhunt decided and why.

What the agent did

The prompt and the response, plus every tool call, data fetch and database query the agent made.

What it touched

Every record, document and file the action read, searched or sent, including data it pulled but didn’t need.

What Darkhunt decided, and why

Allowed, flagged or stopped, with the reason, such as an action that went beyond the task. A tool call that doesn’t fit can be stopped before it runs.

When it happened

Every entry is timestamped for audit.

You can search the record and replay any session.

Who you send it to

The same record answers three different people.

Your customer

“Can we trust your agent with our data?”

Show them an independent check on what your agent does, instead of asking them to trust how you built it.

Your security team

“What did the agent do last Tuesday?”

Search the record and replay what happened step by step.

Your auditor or regulator

“Show me that every agent stayed within its task.”

Filter the record by framework and send the matching decisions as evidence.

Mapped to the frameworks your auditors use

Each decision in the record is tagged by framework, so you can answer an audit question without rebuilding the evidence by hand.

Visit the Trust Center ↗

Framework

What Darkhunt maps

OWASP Top 10 for LLMs

Application security

Flagged and stopped actions tagged by category, such as prompt injection, data leakage or excessive agency.

NIST AI RMF

Risk framework

Activity mapped across Govern, Map, Measure and Manage.

EU AI Act

Regulation

Mapping in progress, planned for 2027.

MITRE ATLAS

Threat matrix

Stopped actions tagged with the matching ATLAS technique, such as prompt injection.

ISO 42001

AI management

Decisions mapped to the ISO 42001 controls for AI management systems.

HIPAA

Healthcare

Every access to patient data and every disclosure recorded, with what was sent and why, to support HIPAA audits and minimum-necessary reviews.

Run Darkhunt in your own infrastructure and the record stays there too.

See deployment and compliance