Security audit

Kim Msg Account Skill

Security checks across malware telemetry and agentic risk

Overview

This is a coherent Kim/OpenClaw setup helper, but it asks for and stores real messaging credentials in ways that could expose them.

Review before installing. Use only on a trusted workstation, avoid pasting appKey, secretKey, or verificationToken into chat, prefer a secure secret store or local masked input, check the @ks-openclaw/kim package source/version, and expect the setup to modify OpenClaw configuration and possibly restart the gateway.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Intent-Code Divergence

Medium

Confidence: 97% confidence
Finding: The README correctly warns that appKey and secretKey must not be exposed in code or shell scripts, but then immediately instructs users to set them via command-line arguments and store them in a plaintext JSON config file. This creates a real secret-handling weakness because secrets entered on the command line may be exposed through shell history, process inspection, logs, or copied documentation, and plaintext local config storage increases the chance of accidental disclosure.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill asks users to provide sensitive secrets including appKey, secretKey, and verificationToken so it can 'automatically configure' the system, but it does not warn users about how those credentials will be handled, stored, masked, or protected. This is dangerous because it normalizes sharing production credentials into a conversational channel, increasing the risk of credential leakage through chat logs, agent memory, telemetry, or operator access.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The verification and troubleshooting guidance includes commands like 'openclaw config get channels.kim.secretKey' and broader config retrieval that may print stored secrets in plaintext. Exposing secrets during routine validation or support greatly increases the chance of accidental disclosure via terminal history, screen sharing, logs, copied output, or chat transcripts.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The script interactively collects appKey, secretKey, and verificationToken, then persists them with `openclaw config set` without any warning about how those values will be stored, who can read them, or whether shell history/logging may expose them. In a setup assistant for an IM integration, this increases the chance that operators handle secrets insecurely or store them in plaintext configuration accessible to other local users or support tooling.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The script prompts for appKey, secretKey, and verificationToken, then persists them via `openclaw config set` without warning the user that these secrets may be stored on disk, retained in local configuration, or exposed to other local users/processes depending on OpenClaw's config handling. In a setup helper for an IM integration, this increases the chance of accidental credential exposure and weak secret hygiene even if there is no direct exfiltration in the script itself.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.