ARK is a command-and-control (C2) framework for authorized offensive security work: red team engagements, penetration tests, CTF/lab environments (e.g. Hack The Box), and defensive research.

You must have explicit written permission (or own the systems) before deploying implants, listeners, or automated agents against any target. Misuse may violate criminal and civil law. The maintainers accept no liability for unlawful or unauthorized use.

This project is early-stage. Treat it as lab-grade, not a hardened enterprise product.

Please report security issues privately so we can fix them before public disclosure.

1. Preferred: Open a private security advisory on GitHub for KKingZero/ARK if available, or email the repository owner via the contact listed on their GitHub profile.
2. Include:
   • Affected component (teamserver, Go implant, C implant, operator, AI agent)
   • Version / commit hash
   • Reproduction steps
   • Impact (RCE, auth bypass, crypto weakness, DoS, data leak, etc.)
3. Allow a reasonable window for a fix before public write-ups (we aim to acknowledge within 7 days and ship a fix or mitigation when practical).

Do not file public GitHub issues for unfixed vulnerabilities that could be abused against third parties.

• Authentication/authorization bypass on the gRPC operator API or mTLS seats
• Implant auth (HMAC), session crypto (AES-GCM), or replay weaknesses
• Path traversal / arbitrary file access beyond intended task jails
• Unauthenticated resource exhaustion on listeners (HTTPS/DNS)
• Privilege issues in approval dual-control (self-approve, missing identity)
• Unsafe defaults that enable high-risk auto-execution without operator intent

• “The tool can be used maliciously” (it is a C2; authorized use is the control)
• Social engineering of operators
• Vulnerabilities only in third-party dependencies with no ARK-specific impact (still welcome if you note the package and version)
• Issues requiring already-compromised teamserver credentials and no additional privilege gain

Operators should:

• Generate unique implant secrets per build; never commit real secrets
• Keep ~/.ark/ (certs, DB, llm.yaml API keys) private (0600 / 0700)
• Use dual-control approver seats for high-risk tasks in multi-operator settings
• Disable or carefully scope auto-harvest in shared environments
• Point listeners only at authorized lab/engagement infrastructure

We prefer coordinated disclosure. Credit is given to reporters who request it unless anonymity is preferred.