CHANGE → CONSEQUENCE → REVIEW
Follow one change. Take the evidence with you.
Analyze a fictional change, inspect its connections, and download a one-page summary with the full evidence package. No sign-in required.
A decision-support sandbox. All examples are synthetic. Results are deterministic estimates, not compliance findings or deployment authorization. A qualified human must review the evidence.
01 / DEFINE THE CHANGE
Select a scenario
Quality tests alter a copy of a built-in scenario. Imported packages use their own evidence. A flagged path remains visible; its priority score does not become a claim of certainty.
Loading the evaluation catalog.
Bring a synthetic input package or AWS Config event
Upload a JSON input package, or an AWS Config event to evaluate against the selected scenario. Maximum 250 KB. Do not upload CUI, operational, personal, or classified information. No connection to an AWS account is made. Imported events clear the example authorization comparison.
02 / TRACE THE CONSEQUENCES
Impact assessment
Assessment summary
RULE-BASED MODELHow to interpret this score
Score = severity weight × 20 + mission weight × 8 + proximity contribution, capped at 100. Severity and mission weights range from 1 to 4. Proximity is max(0, 12 − 2 × dependency hops). A score of 70 or more flags prioritized review. This is an illustrative rule, not a probability of mission failure.
The model follows supplied directed relationships and shows one shortest path per affected node. Missing or incorrect dependencies can change the result. Hashes detect changes when compared with a trusted copy; they do not establish source authenticity or compliance.
REVIEW READINESS / CHECK THE ASSUMPTIONS
What is missing from this explanation?
Focus changes this review desk only. The full graph, other affected systems and complete export stay visible.
Inspect mapping sources, age and disputes
Review-check scope and limitations
Missing data is not proof of safety. A supplied source reference does not authenticate a relationship. This check never grants authorization and does not discover your environment.
One action list for the handoff
Assign a proposed contact, request evidence, or challenge a mapping. These are unverified local annotations, not notifications or closed tickets. Save on this device or download to retain them. Repeating the same local save updates that assessment.
DEPENDENCY PROPAGATION
Follow the impact
Links show the input model’s dependencies. Scores reflect this run’s assumptions.
Inspect dependency paths and model
Systems & missions
IMPACTControls & authorizations
REVIEWBEFORE → AFTER
Authorization & control changes
Scoring provenance & accountable owners
Provenance
Owners
03 / APPLY HUMAN JUDGMENT
Review the evidence
Analyze a scenario to add local review notes. They do not authenticate your identity or authorize a system.
Available after analysis. Your rationale is required; the original model result remains unchanged.
04 / DOWNLOAD THE EVIDENCE
Take a reviewable evidence pack.
A one-page PDF, a plain-language review brief, an action-register CSV and full JSON evidence. Includes missing mappings, path evidence checks and local follow-up notes. The human disposition block stays blank.
No account required. Downloads are created in your browser. Analysis uses the hosted API.
Saved on this device
Explicit local saves only, up to 20 cases. Available to other people using this browser profile. Clear them here or in your browser settings.
Optional authenticated case history
Existing server records remain protected. Authentication is only needed for private server storage and attributed review records.
Access existing private casesPRIVATE WORKSPACE
Saved cases
No authenticated case history is loaded. Anonymous analysis, local notes and downloads are available.