Malware Detected

Potentially dangerous code patterns found in content

Blocked

What Was Found

Our security scanner detected code patterns commonly associated with malware, exploits, or malicious scripts. This content has been blocked and will not be imported into your agent's memory.

Why This Matters

Malware in agent memory could be executed if your agent processes or references this content. Even if the code appears harmless, obfuscated patterns often hide dangerous payloads.

This is a hard block. Content matching malware signatures cannot be imported, even through manual review. This protects both your agent and the systems it interacts with.

Common Causes (May Be False Positive)

Code snippets containing eval() with encoded strings
Shell commands with pipes to execution (curl | bash)
Reverse shell patterns or network backdoors
SQL injection payloads or Log4j exploit strings
Obfuscated JavaScript or PowerShell scripts

What You Can Do

If you believe this is a false positive:

Review the flagged content carefully — is it actually code you trust?
If it's educational content about malware, consider rewriting it to describe the patterns without including executable code

Red team or security research? If you and your agent work on penetration testing, malware analysis, or security research, email us at [email protected] to request an exception. Include your use case and we'll work with you.

Do not attempt to obfuscate content to bypass this check. The scanner will catch most evasion attempts, and bypassing security measures puts your agent at risk.

Review Quarantine Back to Dashboard