Skip to content

Security: Shadowprofessor/DDOS-Attack-Detection

Security

SECURITY.md

Security policy

This is a defensive research project (DDoS detection and human-approved response). Reports about weaknesses in this project's code are welcome.

Supported versions

Only the latest release on main gets fixes. See releases.

How to report a vulnerability

Please do not open a public issue with details.

  1. Preferred: use GitHub private vulnerability reporting on the Security tab → "Report a vulnerability", if the button is shown.
  2. Otherwise: open an issue titled "Security contact request" with no technical details. The maintainer will reply with a private channel.

Please include:

  • the affected file or endpoint;
  • the version or commit;
  • the steps to reproduce against a local instance;
  • the impact.

What to expect:

  • an acknowledgement within 72 hours;
  • a triage assessment with a severity (CVSS) within 7 days;
  • a fix in the next release, credited in the CHANGELOG unless you prefer to stay anonymous.

Severity guidance:

  • Any path that bypasses the human approval gate is critical.
  • Prompt injection that changes a proposed action is high.

Please do not include working exploits against third-party systems.

In scope

  • Bypassing the approval gate: any way a model, LLM output or API call enforces a block without PendingActionManager.approve() by an admin.
  • Prompt injection through telemetry that changes the analyst's output beyond the code-enforced policy.
  • Auth, session, CSRF or RBAC flaws in the API or the SOC web app (/app).
  • Lab safety escapes: traffic leaving the Mininet scope, or the enforcer applying a rule it should refuse.
  • Secrets exposure (for example, the alert webhook URL or API keys appearing in logs or responses).

Out of scope

  • Denial-of-service against your own running instance (it is a lab tool).
  • The mocked Cloudflare connector.
  • Findings that need an already-compromised host or admin key.
  • Reports generated only by automated scanners, without a working reproduction.

Rules

Test only against your own local instance and lab. Never test against systems you do not own. Do not generate attack traffic outside an isolated lab: see the lab safety rules in docs/purple_team/README.md.

There aren't any published security advisories