According to Cursor's advisory (Aug 2, 2025), Cursor allowed writing in-workspace files without approval; editing an existing dotfile needed approval but creating a new one did not. [1][2]
Impact
Chained with an indirect prompt injection, an attacker could make the agent write .cursor/mcp.json and trigger remote code execution without user approval (CVSS 8.5). [1][2]
What was done
Fixed in 1.3.9: the agent is blocked from writing MCP-sensitive files without approval. [1][2]