EchoLeak Zero-Click Prompt Injection on M365 Copilot
Microsoft 365 Copilot
Researchers uncovered a zero-click prompt injection vulnerability in Microsoft 365 Copilot exploiting RAG context. Malicious emails triggered unauthorized data retrieval and subsequent exfiltration without user interaction.
- first report
- 30 Jun 2026
- last update
- 30 Jun 2026
Timeline · 4
- 30 Jun 2026EchoLeak: Zero-Click Prompt Injection Targeting M365 Copilot for Data Exfiltration
- 30 Jun 2026EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM System
- 30 Jun 2026CVE-2025-32711
- 30 Jun 2026Breaking down 'EchoLeak', the First Zero-Click AI Vulnerability Enabling Data Exfiltration from Microsoft 365 Copilot
Sources · 4
News and write-ups
- 2026-06-30CVE-2025-32711cve.org · cve.org/CVERecord?id=CVE-2025-32711
- 2026-06-30Breaking down 'EchoLeak', the First Zero-Click AI Vulnerability Enabling Data Exfiltration from Microsoft 365 Copilotcatonetworks.com · catonetworks.com/blog/breaking-down-echoleak
Research and catalogues
- 2026-06-30EchoLeak: Zero-Click Prompt Injection Targeting M365 Copilot for Data ExfiltrationMITRE ATLAS · atlas.mitre.org/studies/AML.CS0059
- 2026-06-30EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM SystemarXiv · arxiv.org/abs/2509.10540
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 hijackData leak
- severity
- 1 · Research / proof of concept
- targets
- Microsoft
- owasp llm top 10
- LLM01:2025 Prompt InjectionLLM02:2025 Sensitive Information DisclosureLLM06:2025 Excessive Agency
- dates
- happened May 2025 (approx.) · first reported 30 Jun 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
# EchoLeak Zero-Click Prompt Injection on M365 Copilot https://gcdradar.vercel.app/incidents/echoleak-zero-click-prompt-injection-on-m365-copilot · AIR-2026-0356 · closed 2026-07-14 - Lane: Research - Severity: 1 (Research / proof of concept) - Fix status: Fix status unknown - First reported: 2026-06-30 (happened 2025-05) - AI role: Attack on AI - Products: Microsoft 365 Copilot ## Summary Researchers uncovered a zero-click prompt injection vulnerability in Microsoft 365 Copilot exploiting RAG context. Malicious emails triggered unauthorized data retrieval and subsequent exfiltration without user interaction. ## What to do No action recorded; read the sources below and the vendor's advisory. ## Latest updates - 2026-06-30: Breaking down 'EchoLeak', the First Zero-Click AI Vulnerability Enabling Data Exfiltration from Microsoft 365 Copilot (https://catonetworks.com/blog/breaking-down-echoleak) - 2026-06-30: CVE-2025-32711 (https://cve.org/CVERecord?id=CVE-2025-32711) - 2026-06-30: EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM System (https://arxiv.org/abs/2509.10540) - 2026-06-30: EchoLeak: Zero-Click Prompt Injection Targeting M365 Copilot for Data Exfiltration (https://atlas.mitre.org/studies/AML.CS0059) ## Sources - 2026-06-30 MITRE ATLAS: EchoLeak: Zero-Click Prompt Injection Targeting M365 Copilot for Data Exfiltration https://atlas.mitre.org/studies/AML.CS0059 - 2026-06-30 arXiv: EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM System https://arxiv.org/abs/2509.10540 - 2026-06-30 cve.org: CVE-2025-32711 https://cve.org/CVERecord?id=CVE-2025-32711 - 2026-06-30 catonetworks.com: Breaking down 'EchoLeak', the First Zero-Click AI Vulnerability Enabling Data Exfiltration from Microsoft 365 Copilot https://catonetworks.com/blog/breaking-down-echoleak Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.
Related
- closedLenovo AI Chatbot Lena Cross-Site Scripting via Prompt Manipulation in Lenovo AI Chatbot
Research · since 18 Aug 2025
AI - closedCopilot Studio Data Exfiltration via Prompt Injections in Copilot Studio Agent
Research · since Jun 2025 (approx.)
AI - closedClaude Code GitHub Action Claude Code GitHub Action Secret Exposure via Prompt Injection
Research · since 5 Jun 2026
AI - closedMCP SDKs Credential Exfiltration via Remote Poisoned MCP Tool
Research · since 1 Apr 2025
AI - closedCursor Data Exfiltration via MCP Server Prompt Injection Targeting Cursor
Research · since 24 Jun 2025
AI - closedMicrosoft 365 Copilot Microsoft 365 Copilot Financial Transaction Hijacking via RAG
Research · since 8 Aug 2024
AI