Prompt-Based Attacks Against Gemini via Calendar Invitations
Gemini
Researchers demonstrated how malicious instructions embedded in Google Calendar invitations and shared files could influence Gemini assistants during retrieval. The indirect prompt injection caused the model to misuse authorized tools.
- first report
- 31 Jul 2026
- last update
- 31 Jul 2026
Timeline · 2
Sources · 2
News and write-ups
- 2026-07-31Invitation Is All You Need: Hacking Geminisafebreach.com · safebreach.com/blog/invitation-is-all-you-need-hacking-gemini
Research and catalogues
- 2026-07-31Prompt-Based Attacks Against Gemini via Calendar InvitationsMITRE ATLAS · atlas.mitre.org/studies/AML.CS0063
Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.
Classification
- lane
- Research
- AI role
- Attack on AI
- categories
- Prompt injectionAgent hijack
- severity
- 1 · Research / proof of concept
- targets
- Google, Enterprise deployment
- owasp llm top 10
- LLM01:2025 Prompt InjectionLLM06:2025 Excessive Agency
- dates
- happened 6 Aug 2025 · first reported 31 Jul 2026
- origin
- MITRE ATLASDerived from MITRE ATLAS (Apache 2.0, © MITRE).
- record
- confidence 0.95 · created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z
Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned
Copy a brief
This record as a short Markdown brief, for a ticket, a chat or a coding agent.
the brief as text
# Prompt-Based Attacks Against Gemini via Calendar Invitations https://gcdradar.vercel.app/incidents/prompt-based-attacks-against-gemini-via-calendar-invitations · AIR-2026-0361 · closed 2026-08-14 - Lane: Research - Severity: 1 (Research / proof of concept) - Fix status: Fix status unknown - First reported: 2026-07-31 (happened 2025-08-06) - AI role: Attack on AI - Products: Gemini ## Summary Researchers demonstrated how malicious instructions embedded in Google Calendar invitations and shared files could influence Gemini assistants during retrieval. The indirect prompt injection caused the model to misuse authorized tools. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-07-31: Invitation Is All You Need: Hacking Gemini (https://safebreach.com/blog/invitation-is-all-you-need-hacking-gemini) - 2026-07-31: Prompt-Based Attacks Against Gemini via Calendar Invitations (https://atlas.mitre.org/studies/AML.CS0063) ## Sources - 2026-07-31 MITRE ATLAS: Prompt-Based Attacks Against Gemini via Calendar Invitations https://atlas.mitre.org/studies/AML.CS0063 - 2026-07-31 safebreach.com: Invitation Is All You Need: Hacking Gemini https://safebreach.com/blog/invitation-is-all-you-need-hacking-gemini Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- liveGemini Adaptive Prompt Injection Attacks Scaled Using Agent Swarms
Research · since 10 Nov 2026
AI - closedGemini Google Gemini Delayed Tool Invocation Bypass
Research · since Feb 2024 (approx.)
AI - liveMobileProbe: Adaptive Security Evaluation of Mobile Agents via Trajectory-Aware Injection
Research · since 13 Nov 2026
AI - livePrompt Injection, Esq.: pwning AI agent guardrails with lawyerspeak
Research · since 11 Nov 2026
AI - liveLLM agents Defending LLM Agents Against Parameter Poisoning
Research · since 4 Nov 2026
AI - liveLLM search agents Adversarial Self-Evolving Attacks on Search Agents
Research · since 4 Nov 2026
AI