How long does the audit take?

Most audits are completed within 5 business days of kickoff. You receive the written findings report + a 30-minute walkthrough call.

Do I need to give you access to our systems?

No direct access required. You share your system prompts, agent configurations, and API setup. The audit is a review of your AI stack's security posture, not a penetration test of your infrastructure.

AI Security Audit

Find what your red team missed before attackers do.

A comprehensive review of your AI stack for prompt injection risks, data leakage, and insecure agent configurations. Delivered as a written findings report with prioritised remediation steps.

$6,000 flat fee · 5 business days

Get Your Audit →

or book a free 15-min call first

What's included — 60 second walkthrough

Which service do I need?

$6K · Audit

You want to know what's wrong. A one-time review with a prioritised fix list. Start here.

$8K · Guardrails MOST CHOSEN

You know the gaps and want them closed. I implement + test the guardrails in your stack.

$12K · Red Team

You need proof. Adversarial testing + a board-ready compliance report (EU AI Act, NIST).

What's tested

Prompt injection surface mapping — identify every path an attacker can manipulate your LLM's instructions
Model output & data leakage review — catch PII, secrets, and internal data bleeding through responses
API key & credential exposure scan — find keys leaking in prompts, logs, or agent tool access
Agent tool permission review — verify agents can't access tools or data they shouldn't
Written report + remediation roadmap — prioritised findings with fix recommendations

Mapped to OWASP LLM Top 10

Every finding is traced to the OWASP Top 10 for LLM Applications (2025) so your team and regulators speak the same language:

LLM01 — Prompt Injection — direct & indirect instruction overrides via retrieved content
LLM02 — Sensitive Information Disclosure — PII, secrets, and training-data leakage through responses
LLM03 — Supply Chain — vulnerable plugins, packages, and third-party model dependencies
LLM05 — Improper Output Handling — XSS / injection via unfiltered model output
LLM06 — Excessive Agency — dangerous tool/function permissions granted to autonomous agents
LLM08 — Excessive Permission — over-broad access granted to the model or its tools

Full coverage matrix (LLM01–LLM10) is in the written report. NIST AI RMF and EU AI Act articles are cross-referenced where relevant.

What you get

A written findings report (PDF) with:

Every vulnerability found, ranked by severity
Specific remediation steps for each finding
A 30-minute walkthrough call with Mathew Gale
Clear next steps — fix it yourself or engage GaleOps for Guardrail Setup

How it works

Day 1: You share your system prompts, agent configs, and API setup
Days 2–4: I audit your stack against real-world attack patterns
Day 5: Written report delivered + walkthrough call

Frequently asked

What does the audit include?

Prompt injection surface mapping, model output & data leakage review, API key & credential exposure scan, agent tool permission review, and a written report with prioritised remediation roadmap.

How long does it take?

Most audits are completed within 5 business days. You get the written report + a 30-minute walkthrough call.

Do I need to give you system access?

No direct access required. You share prompts, agent configurations, and API setup. It's a security posture review, not infra pen-testing.

What happens after?

You get the report with prioritised findings. If you want GaleOps to fix the issues, the AI Guardrail Setup ($8,000) is the natural next step.

Ready to find your gaps?

$6,000 flat. Written report in 5 days. No retainers, no surprises.

Get Your Audit →

or book a free 15-min call first

Want the full stack?

After the audit, fix everything with Guardrails, keep watch with a retainer, or go deeper with a Red Team engagement.

Continuous AI Security

$2,500/mo

Monthly regression testing, guardrail drift checks, updated technical docs

Talk to Mathew →

AI Guardrail Setup

$8,000

Design + implement production guardrails

Learn more →

AI Red Team & Compliance

$12,000

Adversarial testing + EU AI Act / NIST alignment

Learn more →