SearchLeak and a three-CVE LiteLLM chain broke the same AI trust boundary in two weeks. A 5-check audit maps each gap to a CVE, a verify command, and a fix.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results