Plannix Bot security
Security is the foundation of Plannix Bot. Read how our defense-in-depth architecture ensures safe execution of untrusted code, prevents unauthorized data egress, and satisfies stringent enterprise compliance standards.
Core security principles
Zero Trust Architecture
Every tool invocation, API call, and memory access is authenticated and authorized independently.
Hardware MicroVMs
KVM-based microVMs provide independent Linux kernels, preventing container breakout vulnerabilities.
Encrypted at Rest
Customer-managed KMS keys (AWS KMS, GCP Cloud KMS, Azure Key Vault) with envelope encryption.
Independent Audits
Annual SOC 2 Type II examinations and continuous third-party penetration testing.
Multi-layered sandbox isolation
Traditional Docker containers share the host kernel, leaving them vulnerable to privilege escalation bugs. Plannix executes bot tasks inside hardware-virtualized microVMs (powered by Firecracker and gVisor). Each microVM has its own isolated memory, virtual disk, and seccomp filters.
Egress controls & network firewalls
By default, sandbox network access is restricted to an allowlist of approved package repositories and developer APIs. Attempts to connect to unrecognized IP addresses or ports are blocked and flagged for administrator review.
Data protection & encryption
All customer data in transit is protected using TLS 1.3 with forward secrecy. At rest, ephemeral disks use volatile cryptographic keys that are destroyed immediately upon task termination.
Vulnerability management & pen tests
We maintain a continuous vulnerability disclosure and bounty program, accompanied by quarterly grey-box penetration tests conducted by accredited cybersecurity firms.