KVcore MCP CLI

• Analyzed: May 12, 2026, 4:04 AM
• Findings: 1
• Version: 0.1.0
• Hash: 9b15de70f5c658ea8b1818dcacfe28c06301ee387fba6d07114abb75d60151b2
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill is classified as suspicious due to several high-risk capabilities described in `SKILL.md`. It explicitly offers 'raw endpoint access' via `kvcore_request`, allowing the AI agent to make arbitrary API calls to the KVcore system. Furthermore, it allows specifying a custom `KVCORE_BASE_URL`, which could redirect sensitive API interactions to an attacker-controlled server. The `twilio_call_create` function also permits arbitrary TwiML, enabling potential social engineering or phishing calls. These features, while potentially intended for flexibility, introduce significant prompt injection vulnerabilities, allowing an attacker to coerce the AI agent into performing unauthorized data exfiltration, modifications, or other harmful actions against the KVcore API or via Twilio.