ClawHub
Back to skill

Security audit

dmp-cli

Security checks across malware telemetry and agentic risk

Overview

This skill is coherent for operating a DMP CLI, but it deserves review because it normalizes sensitive audience-data operations, plaintext credentials, and unverified binary installation with limited safety guidance.

Install only if you trust the Mingdata DMP CLI source and can verify the release binary independently. Use a secret manager or CI secret store instead of pasting plaintext secrets into commands, avoid logging environments or audience file paths, restrict permissions on local DMP config, and confirm you are authorized before uploading, transforming, syncing, creating, or modifying audience and deal data.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill explicitly instructs users to provide `DMP_SECRET_KEY` in plaintext and notes that credentials may be stored in `~/.dmp/config.yaml`, but it does not warn about shell history, process-environment leakage, CI log exposure, or persistent local storage of secrets. In a security-sensitive CLI skill, documenting secret handling without clear safeguards can lead users or agents to expose credentials unintentionally.

Missing User Warnings

Medium
Confidence
82% confidence
Finding
The examples show handling of upload credentials and audience data flows, including secrets via environment variables and paths to potentially sensitive datasets, but provide no warning about secure secret handling, least-privilege access, or privacy controls. In a DMP context, these commands operate on audience identifiers and uploaded user data, so omission of handling guidance can lead to credential leakage, misuse of production data, or privacy violations by operators copying examples directly.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The installation instructions fetch release metadata and a binary from the network, mark it executable, and move it into /usr/local/bin with sudo, but provide no integrity verification, signature/checksum validation, or trust warning. If the release source, network path, or fetched artifact is tampered with, users could install and execute a malicious binary with elevated trust on their system.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.