Claude agents negotiated 186 deals and caused employees to exchange physical goods
Review RAT-137 · Catalogue RAT-137
Anthropic explicitly authorized employee agents to transact autonomously in Project Deal using supplied budgets. The physical exchange of goods is intended experiment behavior, not a security boundary violation.
Reviewed: · Rule: roguewatch-organizational-harm-v1
Dates and discovery
- Incident date
- Not disclosed
- Added to research / record
- Posted to RogueWatch
- No linked public incident
- First observed
- 2026-10-30T13:13:39.761834Z UTC
- Discovery channel
- User-supplied comparison audit
- Discovery source
- https://rogueaitracker.com/incidents/claude-agents-negotiated-186-deals-and-caused-employees/
Dates use America/Los_Angeles. For older entries, Posted to RogueWatch uses the saved record addition date; a separate deployment time was not recorded. Future posting timestamps record first public qualification. Source publication and incident occurrence are separate.
Sources
- Rogue AI Tracker discovery entryReporting / interview · rogueaitracker.com
- www.anthropic.comOriginal account · www.anthropic.com
Decision history
- · Case-by-case comparison audit
RAT-137: out-of-scope. Anthropic explicitly authorized employee agents to transact autonomously in Project Deal using supplied budgets. The physical exchange of goods is intended experiment behavior, not a security boundary violation. Source checks: read: https://www.anthropic.com/features/project-deal. Anthropic explicitly authorized employee agents to transact autonomously in Project Deal using supplied budgets. The physical exchange of goods is intended experiment behavior, not a security boundary violation.
Evidence source
A review decision applies only to the named scope. Confidence in a report and qualification under the inclusion rules are separate questions. Review records do not add incidents to the globe or leaderboard.