Skip to content

Security: PSaiSurya/Ankusha

Security

SECURITY.md

The Mahout's Creed

Philosophy: Security through Minimalism (O(1) Silence)

Ankusha is engineered as a Native HPC Manifestation. In Critical Infrastructure environments, the most significant vulnerability is often the "Security Tooling" itself. By eliminating the need for a web server, a database, or third-party interpreters (Python/Node.js), Ankusha reduces the Supply Chain Attack Surface to absolute zero. We follow the principle of "Living off the Land": if the binary isn't already part of the LINUX core or the Slurm environment, Ankusha does not use it.

Threat Model & Mitigation Strategy

Threat Vector Ankusha Defensive Control
Privilege Escalation Zero-Sudo Architecture. Ankusha is a user-space Mahout. It never requests or requires root access. It operates strictly within the permissions of the current user, ensuring it cannot be used as a vector for privilege escalation.
Supply Chain Compromise Dependency-Free Execution. By eschewing pip, npm, and external modules, Ankusha is immune to "Dependency Confusion" or "Typosquatting" attacks. The code is 100% auditable Bash.
Remote Code Execution No Network Footprint. Ankusha opens zero sockets, starts no listeners, and has no "phone home" telemetry. It is fundamentally invisible to network-based scanners (Nmap/Zmap) and remote exploit attempts.
Lateral Movement Local Scoping. Ankusha does not store or transmit credentials. It leverages existing Slurm environment tokens for data fetching, preventing the storage of "secrets" in memory or on disk.
Data Integrity Read-Only Manifestation. Ankusha treats system data (via /proc and sinfo) as read-only. It does not modify system states, ensuring no interference with the operational integrity of the mammoth (the HPC).

Secure Operation Practices

1. User-Space Resilience

Ankusha is designed for high-security, air-gapped environments. It fetches metrics directly from standard Slurm binaries and the /proc filesystem. Because it runs as a standard process, it is subject to the same SELinux or AppArmor profiles as any other user job, ensuring a "Sandbox-by-Default" posture.

2. Environment Sanitization

  • Variable Isolation: All script logic uses local-scoping to prevent sensitive environment variables from leaking into the global shell.
  • Input Neutralization: Ankusha is a non-interactive dashboard. It does not accept external flags or user-defined strings during execution, neutralizing the risk of Shell Injection.
  • Zero Persistance: No job metadata or cluster telemetry is written to persistent logs. The "Manifestation" exists only in volatile memory while the dashboard is active.

Reporting a Vulnerability

As a "User-space Mahout," I prioritize the logic of control. If you identify an edge case where the tool could cause resource exhaustion or bypass Slurm's standard user-level restrictions, please report it via GitHub Issues. Logic-based bypasses are treated with the highest priority.

There aren't any published security advisories