Cassandra by
Oct 29 · Claude Code 2.1.313 shipped with an SSH key, a LaunchAgent and a tunnel. Cassandra checks for all three.

Are my AI tools backdoored?

One command checks every extension in your editors and browsers against 10,990 known-malicious ones, looks each one up on its marketplace, and looks for the traces of recent attacks. Free, read-only, about ten seconds.

Cassandra · report Example
Ran the check on macOS Done
Editor extensions
31 · Cursor, VS Code
Browser extensions
47 · Chrome, Brave
AI tools
claude, cursor-agent
API key variable names
2, names only
Matched 78 extensions against 10,990 known-malicious ones Clear

Lists: Microsoft's VS Code Marketplace removals, and chrome-mal-ids by The Privacy Commons Institute. Refreshed every 6 hours.

Looked for traces of the Claude Code 2.1.313 backdoor 3 found
  • Attacker SSH key (svc-telemetry) in authorized_keys
    Found
  • Fake Apple LaunchAgent com.apple.sysmond.plist
    Found
  • Reverse-tunnel service com.apple.sysmond loaded
    Found
  • Live connection to the attacker server 134.209.82.225
    Clear
Checked 31 editor extensions on their marketplaces Clear

Looked up on the VS Code Marketplace and Open VSX for pulled versions and look-alikes.

Critical Compromised

Attack traces found on this machine (3).

Something known to be malicious is on this machine. Treat it as compromised.

Your report link arrives by email. Read the script before you run it, and run it only on a machine you're allowed to check.

Already ran a check?

How it works

  1. 01

    Enter your email

    You get a one-time command for this check. It works once and expires in a day.

  2. 02

    Run it in a terminal

    It lists your extensions and looks for known attack traces, prints every fact, and asks before sending anything.

  3. 03

    See your risk level

    Here and in your terminal. A link arrives by email; it signs you in to the full report.

curl -fsSL https://ciphero.ai/cassandra/scan/YOUR-CHECK-TOKEN | bash
irm "https://ciphero.ai/cassandra/scan/YOUR-CHECK-TOKEN?format=ps1" | iex

Your own token replaces YOUR-CHECK-TOKEN · single use · no sudo · Read the script

◆ CASSANDRA  malicious extension check · by Ciphero

Editor extensions   31 found, ids and versions only
Browser extensions  47 found in chrome brave, ids and versions only

Claude Code 2.1.313  incident traces
○ Attacker SSH key svc-telemetry in authorized_keys clear
○ Attacker SSH key fingerprint in authorized_keys clear
○ Fake Apple LaunchAgent com.apple.sysmond.plist clear
○ Reverse-tunnel service com.apple.sysmond loaded clear
○ Tunnel files ~/.config/.smt and ~/.config/.smtk clear
○ Staging files .stk_* in the temp directory clear
○ Live connection to the attacker server 134.209.82.225 clear

AI tools  claude, cursor-agent
Key variable names  ANTHROPIC_API_KEY, OPENAI_API_KEY

Ready to send 88 facts to Ciphero. No file contents, no key values.
Send them? [Y]es / [n]o / [s]how every line:

What it reads

  • Extension ids and versions in Cursor, VS Code, VS Code Insiders, Windsurf, VSCodium, Kiro and Trae
  • Extension ids and versions in Chrome, Edge, Brave, Arc, Firefox and other Chromium browsers
  • The signs of specific attacks, such as the Claude Code 2.1.313 backdoor's SSH key, LaunchAgent and tunnel
  • Names of AI tools and API key variables, never their values

What it never does

  • Read or send file contents, keys or secrets
  • Install, schedule or leave anything behind
  • Ask for sudo or admin rights
  • Send anything before you say yes

The verdict comes from fixed rules: public known-malicious lists, marketplace lookups and published attack indicators. Cassandra's written analysis explains the result; it can't change it.

Cassandra by Ciphero matches your editor and browser extensions against public known-malicious lists and the marketplaces, and looks for the traces of specific incidents. It is not antivirus.

Extension data from Microsoft's VS Code Marketplace removals and chrome-mal-ids by The Privacy Commons Institute, licensed CC BY 4.0.