OKKI CRM

Security checks across malware telemetry and agentic risk

Overview

This CRM skill is purpose-aligned, but it can modify business records and send email through an external client and cached tokens that are not included or fully described.

Review and trust the referenced okki_client.py and config before installing. Use a least-privilege OKKI sandbox account, restrict the #okki channel to authorized users, confirm where tokens are cached, and require explicit human approval for any create, update, delete, order, or email action.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill documents commands and code examples that can create, update, delete CRM records, create orders, and send email, but it does not prominently and consistently warn that these operations cause real remote side effects. In an agent setting, examples often become de facto instructions for autonomous execution, so insufficient disclosure can lead to unintended modification of customer data or outbound communications even if some risk labels are present.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal