vulns.co
/
GKData.io MCP and API Buy me a coffee

Google · AI integration boundaries

Gemini Enterprise connected-content trust failure allowed persistent-memory modification

Gemini Enterprise could treat retrieved Jira content as permission to change saved assistant memory. Using two controlled accounts, the researcher demonstrated deletion of the recipient's memories; this required access to shared-project content and a user-initiated retrieval. The researcher reports a $15,000 Google bounty, but the demonstration does not establish actual customer compromise or broader cross-tenant access.

Read the original writeupUSD 15,000 reported award · Reviewed 2026-10-02

Root cause

Retrieved collaboration content could influence changes to saved memory without the confirmation prompt typically used for other user-data actions. The researcher compared those two paths. Permission to read connected data was being treated as authority to change the user’s lasting assistant state.

Demonstrated impact

In a two-account setup, the researcher confirmed deletion of the test recipient’s saved memories. The scenario required shared-project content access and user-initiated retrieval. Broader cross-tenant access, mailbox compromise and lasting changes to every future response were not demonstrated.

What to check for

  • Model persistent memory as a security-sensitive write operation with its own authorization decision.
  • Preserve provenance when retrieved content passes between a connector, model and state-changing component.
  • Require a trusted authorization signal for each write; permission to summarize content should not imply permission to alter saved preferences.
  • Separate observed state changes from speculative downstream behavior in impact reports; the public article does not establish the vendor’s remediation design.

Award and evidence

USD 15,000Bug Bounty · Researcher Reported
How the award was verified, and what the record does not establish

The researcher reports being paid $15,000 for this distinct finding; do not combine the separate $1,337 example. Exact award/settlement date is unknown. Dollar-denominated Google bounty; USD normalization. The individual write-up uses the $ symbol rather than spelling out USD.

  • No vendor-hosted confirmation of this individual payout found
  • Do not confuse this finding with the separate $1,337 memory issue mentioned in the introduction
  • Remediation date, exact patch design and current status are not independently established
  • The reported demonstration used two researcher-controlled accounts; it does not establish actual customer compromise

Timeline

Published
2026-03-12inferred · DEV article March 12, 2026; the researcher's Reddit write-up is dated March 9, 2026. Original X post is linked but its exact date was not independently established. No exact report, award, or fix dates were found.
Public Disclosure
2026-03-09explicit · Earlier researcher publication on Reddit; linked original X post might be earlier.

Related visual models

Related cases

All 5 cases in AI integration boundaries

Sources

  1. Google paid me $15,000 for this Prompt Injection bug.Behi · reviewed 2026-10-02
  2. Supporting primary disclosure sourceBehi · reviewed 2026-10-02

Record reviewed 2026-10-02. Snapshot a066321c166a. Open the complete JSON contract.

Library synced 2026-11-19 · cases last reviewed 2026-10-05

a066321c166a · JSON exports & schemas · CC BY 4.0 content / MIT software