mongodb-attack

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill embeds literal passwords in examples and shows commands that pass credentials on the command line or inside payloads, which forces the LLM to handle/output secret values verbatim (high exfiltration risk).

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). One GitHub repo (https://github.com/Charlie-belmer/nosqli) is a personal/low-reputation repository distributing a NoSQL testing tool and matches high-risk indicators for GitHub sources from unknown authors, while the other URLs are either a local test endpoint or a well-known project.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is a detailed, actionable offensive guide for exploiting MongoDB (unauthorized access, NoSQL injection, RCE, credential extraction, privilege escalation and data exfiltration), indicating deliberate malicious intent and tools for abuse.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill includes an explicit runtime command that clones and executes remote code from https://github.com/codingo/NoSQLMap (git clone https://github.com/codingo/NoSQLMap; cd NoSQLMap; python3 nosqlmap.py), which fetches and runs external code during execution.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly instructs creating and elevating database users (db.createUser, db.grantRolesToUser), dumping data to /tmp, and reading system configuration (/etc/mongod.conf), i.e. clear state-changing and privilege-escalation actions that could modify the host (notably if TARGET is local).