ClawHub
Back to skill

Security audit

Atlas Admin Console Free

Security checks across malware telemetry and agentic risk

Overview

This skill is mostly an Atlas API helper, but it mixes read-only documentation language with credentialed admin actions that can change MongoDB Atlas resources.

Review before installing. Use this only for MongoDB Atlas Admin API tasks, provide least-privilege Atlas credentials, and require explicit user approval before any POST, PUT, PATCH, or DELETE command. Treat the free-edition export and batch-call claims as inconsistent until the publisher clarifies them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Description-Behavior Mismatch

High
Confidence
95% confidence
Finding
The manifest and summary position the skill as a read-only browser/query tool, but the body explicitly documents POST/PUT/PATCH/DELETE execution paths, including create/modify/delete operations once credentials are configured. This mismatch can mislead users or higher-level agents into invoking a tool they believe is non-mutating, increasing the chance of unintended destructive actions against Atlas resources.

Intent-Code Divergence

Medium
Confidence
78% confidence
Finding
The documentation states that output_format supports export/save/convert operations, while later claiming the free edition does not support exporting results. Contradictory capability statements can cause agents or users to rely on unavailable or differently-behaving output paths, which is dangerous when outputs may include sensitive API responses or credentials-adjacent metadata.

Intent-Code Divergence

Medium
Confidence
76% confidence
Finding
The skill says free edition does not support batch API calls, yet earlier instructs that read-only operations can be executed in bulk. This inconsistency can lead automation to issue repeated or pseudo-batch queries unexpectedly, risking rate-limit exhaustion, noisy monitoring activity, or broader-than-intended data retrieval.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The invocation description includes broad triggers such as database operations, SQL queries, and data storage management, even though the skill is specifically for MongoDB Atlas Admin API browsing/calling. Overbroad routing language can cause the skill to activate for unrelated database tasks, unnecessarily exposing credentials or enabling admin-level API actions in contexts where a narrower tool should be used.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The capability coverage section enumerates a very large, vague keyword set and claims all corresponding scenarios are supported, without meaningful activation constraints. This increases the chance that an agent selects the skill based on superficial keyword matches rather than true fit, which is especially risky because the skill has exec capability and can perform authenticated Atlas operations.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.