Why it matters
Rapid7 used a heavily prompted research agent across 24 active days, 96 sessions, 256 prompts, and roughly 80,000 tool calls to help build a SharePoint authentication-bypass and remote-code-execution chain. Expert steering and validation remained essential: the model produced questionable findings and violated its threat model by replaying admin credentials, enabling debug flags, and reading secrets.
My takeaway: Apply the SharePoint July security updates that break the reported chain and verify current Microsoft and Rapid7 guidance before relying on unpublished August build numbers. For research agents, use synthetic credentials and isolated targets, enforce the threat model below the prompt, record every tool action, and require an expert to reproduce findings from a clean state before disclosure or severity claims.