Make the assessment process as reviewable as the result.
ChaosEngine replaces fragile hand-built pipelines with visible workflows that connect authorized scope, human decisions, worker execution and output artifacts.
This preview represents the current product model. Labels and controls may change during early access; example targets and outputs are illustrative.
Security teams have tools. The missing layer is repeatable execution.
Bash scripts and one-off commands are fast to create, but difficult to review, hand off and reproduce. Tool output becomes fragmented, approval context lives in chat, and the next assessment starts by reconstructing the previous one.
ChaosEngine makes the sequence explicit: what enters the workflow, what can run, who must approve it, where it executes and which artifacts it produces.
Design, govern, execute and inspect from one control plane.
Visual DAG builder
Compose input, tool, script, agent and output nodes on a visible graph with explicit dependencies.
Branch on structured results
Route a downstream node when a JSON field equals, differs from or contains the value your workflow expects.
Edit workflows with AI
Describe a change, review the updated canvas and undo the last AI modification when it is not right.
Model the target precisely
Manage approved targets across domains, IPs, repositories, APIs, mobile apps, networks, files and URLs.
Approval before execution
Move workflows and supporting resources through draft, pending, approved or rejected states with reviewer context.
Catalog reusable building blocks
Keep tool actions, agent definitions and runtime configurations available as governed workflow resources.
Run, schedule, cancel and resume
Track queued and running work, apply node timeouts and retries, cancel an execution or resume from a failed node.
Structured results and artifacts
Keep node status, errors, table-ready outputs and execution artifacts connected to the run that produced them.
Organization controls and MCP
Apply organization-scoped permissions, manage users, retain audit activity and operate supported resources through MCP.
The workflow definition is only the start.
Before execution, ChaosEngine validates the graph and checks the applicable approval state. During execution, nodes run through the configured backend and retain their status and artifacts.
- 01Define authorized scope
Start with a supported target type and the context needed by later nodes.
- 02Compose and validate
Connect nodes into a reachable, acyclic workflow with a clear output path.
- 03Approve
Record the reviewer decision required by the organization and workflow state.
- 04Execute through workers
Run node commands and agent operations in the configured execution environment.
- 05Review artifacts
Inspect node outputs and final exports together with the execution status.
What ChaosEngine is—and is not.
Is ChaosEngine a vulnerability scanner?
No. ChaosEngine orchestrates registered tools, scripts and agents. The scanner remains the tool node; ChaosEngine governs how it participates in the wider assessment.
Is it a managed pentest service?
No. Your team remains responsible for authorization, assessment design, reviewer decisions and interpretation of results.
Is every CLI tool automatically supported?
No. A tool must be represented in the ChaosEngine catalog or deliberately wrapped, and its binary and dependencies must exist on the selected worker.
Does catalog availability prove runtime readiness?
No. Early-access onboarding confirms the actual worker environment and tool behavior before a workflow is treated as executable.
Bring one real assessment workflow.
We are onboarding a focused group of security teams and helping each one translate an existing process into a governed ChaosEngine workflow.