Skip to content

Security: EyeDebugger/eyedebugger

SECURITY.md

Security Policy

Supported versions

There have been no releases yet; fixes land on main. After the first release, the latest minor release gets security fixes.

Report privately

Report vulnerabilities through GitHub private vulnerability reporting: https://github.com/EyeDebugger/eyedebugger/security/advisories/new

Never use a public issue, pull request or discussion. Include the eyedbg version output, your OS/arch, the debug adapter and its version, a reproduction, and the impact.

What to expect

  • The project is volunteer-maintained; we aim to acknowledge within 7 days.
  • We coordinate the fix and disclosure date with the reporter.
  • We credit the reporter unless they decline.

In scope

  • Access control of the daemon IPC endpoint and token file (docs/DESIGN.md §6, §11).
  • Untrusted input: adapter messages, adapter manifests, downloaded adapters and checksum verification (docs/DESIGN.md §7). Your own adapter manifests are trusted like your shell configuration: they name commands eyedbg runs (ADR 0011). eyedbg reading a manifest from anywhere but your config directory (e.g. from a project), or accepting one that fails its permission check (someone else can write it or its directory), is in scope; so is a Python probe or adapter start that imports code from the program's working directory, or running a project venv another user owns or can write.
  • Eval side effects vs. the read-only default (docs/DESIGN.md §11).
  • Redaction and session recordings (docs/DESIGN.md §5, §11).
  • Build, CI and release supply chain.

Out of scope

  • Vulnerabilities in third-party adapters or runtimes (netcoredbg, SharpDbg, debugpy, …): report those upstream, and we'll help coordinate.
  • Attacks that require already executing code as the same OS user as the daemon.

There aren't any published security advisories