Unverified50% confidenceFactExact time
Learna Tutor's output filter (checkTutorOutput) has a dedicated privacy check; when the tutor asks the child for personal information, that response is blocked.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/16/2026
First Seen
Valid until: 1/14/2027
Sources
Related Entities
Related Claims
UnverifiedLearna Tutor's input filter (checkTutorInput) contains eight categories of regex rules: privacy, crisis, bullying, medical/legal, out-of-scope, academic integrity, prompt injection, and unsafe web pages, and rejects unsupported tutor modes.78% similarUnverifiedIn agentic workflows, every tool call, memory read and sub-agent spawn should undergo independent permission checks, rather than verifying the user's identity only once when the workflow starts.63% similarUnverifiedA developer built a benchmark called Unlearning Security Audit in the Kaggle Benchmarking Challenge to test whether large language models can detect disguised unlearning code.63% similarUnverifiedCognition reported that the review step performs best when the reviewing agent does not share context with the coder beforehand.62% similarUnverifiedThe behavioral gate checks whether functionality is actually usable by writing test code that simulates real users using the feature, covering all tests that can be completed without a browser.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1041477API
curl https://kongchang.com/api/v1/knowledge/claims/1041477MCP
get_claim(id=1041477)