The public page shows verified results only. Prompts, rules, configurations and operating procedures remain confidential.
1Scenario 1: AI Offensive (Nativi) vs AIPF
The latest controlled run measured the ability of Oracolo, part of the AIPF line, to detect and block attempts produced by the internal Nativi suite.
All 72 simulations containing attack signals were detected and blocked. The 4 legitimate controls were allowed. Public products remain available on their respective pages. Attack datasets, prompts, configurations, complete evidence and validation procedures remain confidential and can be presented to qualified investors and partners in controlled environments.
2Scenario 2: SlimWire (Context & Token Optimization)
Context optimization benchmark: measures the token and memory savings achieved by applying SlimWire context-compression algorithms on the laboratory's reference files.
3Scenario 3: Vigil Security Proxy (WAF & Integration Tests)
Testing of the Vigil security proxy. Automated tests check network traffic, the fixed rules of the web application firewall (WAF) and the masking of personal data (PII).
Methodological Notes & Isolation
Tests run locally without external services. Evidence and reports are sealed with LogArmor Replay so later changes can be detected. Physical or air-gapped isolation depends on the environment selected by the client.
Confidential validation
MC Project Lab is available to launch pilot projects with companies and industrial partners interested in validating the technology and supporting its development.
MC Project Lab performs authorised tests in controlled environments and delivers a report with verifiable results and evidence. Offensive tools, prompts and operating procedures are not distributed.