CARE CORE V5.8 / EXPLORE
See where the core stops.
Explore recorded checks from the current release. These examples are not a live safety assessment of your situation.
People need protection every time.
Our goal is 100% reliable core values. We have not proved that level of safety across the whole system. This demo must not make life-or-death decisions on its own.
Explore the new core’s checked examples.
Choose a saved test from the V5.8 core. These results were produced by the Python checker, not generated in your browser. No live API call or outside access occurs here.
Passes supplied checks — not permission to execute
The registered host supplied signed evidence for a read within its allowed scope.
Recorded reason: ordered_selection
For a live integration, use API contract 2 or the signed-check MCP tool. Your trusted host must be registered first.
V5.8 / ELEVEN VALUES / CONTRACT 2
Check the action. Check the permission.
Care Core compares what could happen with and without an action. It checks every supplied outcome, then holds requests that conflict with its values or lack permission. It does not understand language or carry out actions.
These are related variations of 16 conditions, not 512 independent attacks. This is not a real-world accuracy score. No new OpenAI comparison or speed benchmark was run for V5.8.
- 01
Exact request
Name the resource, operation, destination and record.
- 02
Evidence source
Check the registered host’s signature and expiry.
- 03
Possible outcomes
Require matching people and situations in the supplied data.
- 04
Values and permission
Reject conflicts and access outside the registered scope.
- 05
Compare options
Prefer less remaining harm; hold unresolved ties.
- 06
Return the result
Explain the decision. Never execute the action.
Signed does not mean true.
A signature identifies a registered source; it does not prove a prediction. Missing people, hidden effects and wrong facts can still mislead this check. Your application must verify actual facts and current permission before acting. Do not use this alone for life-critical decisions.
Separate legacy V5 language-reader demo
LIVE READER: V5 · TEXT ONLY
The demo runs on PROME's development computer, without OpenAI. If that computer is offline, the test cannot finish. PROME can still misunderstand a request even when all steps finish. It cannot control tools or machines. See the current core release.
Try the new reader.
Use an example or edit the request within its listed setting. This is not a general chatbot. Your request is saved privately; do not enter personal or confidential information.
MADE-UP SCENARIO FACTS
A person is in the robot's work area. Inspecting and stopping have low simulated risk; starting has high simulated risk.
Available actions: inspect, start, stop.
03 / KEEP EXPLORING