Autonomous Audits

An autonomous audit lets you brief Golem the way you’d brief a human operator. You write the mission in plain English — the agent plans and executes accordingly.

Property	Value
Default duration	4 hours
Methodology scope	Operator-defined
Post-exploitation	Operator-defined
Best for	Red-team simulations, targeted engagements, novel scenarios

How it works

Instead of choosing from a fixed methodology, you provide a scope — a freeform instruction that’s injected into Golem’s system prompt. The agent treats this as its mission brief and plans accordingly. The scope can be as specific or open-ended as you want:

“Assume breach. Start from a compromised employee laptop. Goal: access production AWS keys.”
“Focus exclusively on the OAuth flow at auth.acme.com. Probe for token leakage, scope confusion, and replay vulnerabilities.”
“Behave like APT29 — slow, low, and persistent. Maximum stealth. Cover tracks.”
“Find any way to access customer PII. Document the full kill chain end to end.”

When to use

Red-team simulations — emulate a specific threat actor’s TTPs
Targeted assessments — test one feature, flow, or component in depth
Assumed-breach scenarios — start from a specific foothold and see where it leads
Novel scenarios — anything that doesn’t fit the shallow/deep templates
Customer engagement — let your security team write the brief, not pick from a menu

What the scope can include

Golem reads the scope as authoritative direction. You can specify:

Element	Example
Starting conditions	”Assume access to employee VPN”
Target focus	”Only the GraphQL API at /api/graphql”
Threat actor profile	”Behave like a financially motivated ransomware affiliate”
Tactical constraints	”No bruteforce. Stealth-only nmap (-T1)“
Success criteria	”Goal: prove exfiltration of customer database”
Out of scope	”Do not touch the WordPress blog at /blog/*“

Running an autonomous audit

API

curl -X POST https://api.usegolem.ai/api/workspaces/{workspaceId}/audits \
  -H "Authorization: Bearer $GOLEM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "autonomous",
    "targets": ["acme.example.com"],
    "scope": "Focus exclusively on the OAuth flow at auth.acme.example.com. Probe for token leakage, scope confusion, redirect_uri bypass, PKCE downgrade, and replay vulnerabilities. Goal: capture an access token belonging to another user."
  }'

Dashboard

Open the workspace
Click New Audit
Select Autonomous mode
Write your scope in the Mission Brief field
Confirm targets
Click Launch

Writing effective scope

Golem responds best to briefs that mirror real operator language: Good scope

Assumed breach. Operator has obtained a low-privilege user
session at app.acme.example.com (Bearer token in $TOKEN env var
of the sandbox). Goal: escalate to admin role and dump the
user table. Out of scope: any subdomain other than app.*

Less effective scope

Please test the application thoroughly.

The more concrete the goal, the more focused the engagement.

Authorization caution

Autonomous mode gives Golem maximum latitude. The scope text you provide is injected into the agent’s system prompt and acts as authoritative direction. Be precise about what’s in scope — and be certain you have written authorization for everything in the brief.

Getting Started

Audits

How It Works

Platform

Billing

Autonomous Audits

Autonomous Audits

How it works

When to use

What the scope can include

Running an autonomous audit

Writing effective scope

Authorization caution

Next steps

Golem Agent

Toolkit

​Autonomous Audits

​How it works

​When to use

​What the scope can include

​Running an autonomous audit

​Writing effective scope

​Authorization caution

​Next steps

Golem Agent

Toolkit

Autonomous Audits

How it works

When to use

What the scope can include

Running an autonomous audit

Writing effective scope

Authorization caution

Next steps