Plango has launched a command center designed to assess and enforce personal security posture for developers using AI coding agents. The tool specifically targets the risks posed by agents like Claude Code, Cursor, and Codex, which have broad permissions to read files, run commands, and communicate with the internet. By setting strict rules on what these agents can access, Plango aims to stop malicious or accidental data leaks before they happen.
Controlling Agent Access and Secrets
The core functionality revolves around granular control over the local environment. Plango allows users to keep sensitive items such as SSH keys, .env files, and Finder-tagged folders completely out of reach for AI agents. It also addresses the common issue of plain-text secrets by moving keys into the macOS Keychain, reducing the attack surface for any tool that might inadvertently read a directory listing.
Monitoring Network and Skills
Beyond file system permissions, the tool monitors every network connection to decide where apps and agents are allowed to talk. This firewall-like behavior helps stop data exfiltration attempts. Additionally, Plango scans skills and plugins for hidden instructions or hooks that could leak work, and it checks projects for libraries with known CVEs, providing a comprehensive audit of the developer's machine including app permissions and startup items.
Pricing and Platform Availability
Plango is available for Mac with Apple silicon (macOS 15 or newer) and Windows 11. The service offers a free tier that includes agent guards for major AI tools, MCP and plugin guards, and basic security checkups on one computer. A paid plan at $4.99 per month unlocks advanced guard models, Plango Chat, and the ability to use the service on any number of computers. New accounts receive a 7-day free trial of the paid features without needing a credit card.
Key Takeaways
- Plango enforces security rules for AI agents like Claude Code and Cursor to prevent unauthorized file access and data exfiltration.
- The tool integrates with macOS Keychain to move secrets out of plain-text .env files, enhancing local security.
- Pricing is simple: a robust free tier for single-user setups and a $4.99/month plan for multi-device usage and advanced features.
The Bottom Line
As AI agents gain more autonomy in our dev environments, a dedicated sandbox and audit tool is no longer a luxury but a necessity for serious builders.