ClawHub
Back to skill

Security audit

Clickhouse Database

Security checks across malware telemetry and agentic risk

Overview

The skill is a straightforward SingleStore database CLI guide; it includes powerful SQL examples but does not hide behavior, persist, or request access beyond the database actions the user chooses to run.

Install only if you want Codex to help operate SingleStore through CLI commands. Use a read-only or least-privilege database role by default, verify the target host/database before running commands, manually approve any INSERT/UPDATE/DELETE or script execution, and avoid exporting sensitive query results to shared temporary paths.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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 (3)

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill prominently documents INSERT, UPDATE, and DELETE operations without a clear safety gate, confirmation requirement, or warning about irreversible data modification. In an agent context, this increases the chance that a user request or ambiguous prompt could lead to destructive production changes against a live database.

Missing User Warnings

Medium
Confidence
85% confidence
Finding
The example writes query results directly to a local file in /tmp without warning about sensitive data exposure, file permission risks, or accidental overwrite. In agent-driven workflows, exporting database contents to local storage can leak regulated or confidential data to shared locations or leave recoverable artifacts behind.

Missing User Warnings

High
Confidence
96% confidence
Finding
The skill recommends executing SQL scripts with --multiquery and input redirection, but provides no warning about multi-statement execution, partial failure handling, or irreversible schema/data changes. This is particularly dangerous in an agent setting because a single script can bundle destructive operations that are hard to review and easy to run against the wrong environment.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.