Security audit

TaskboardAI Skill

Security checks across malware telemetry and agentic risk

Overview

This is a coherent TaskBoardAI Kanban integration, but it can save task context and final summaries into persistent task-board cards.

Install this only if you want TaskBoardAI to retain task details outside the chat. Verify the npm package and MCP server path, avoid saving secrets or sensitive business data in task cards, and remove the inferred-task trigger if you want cards created only after explicit user commands.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger condition is broadened from an explicit user keyword to any situation where the agent 'determines a task needs tracking'. That creates an authorization ambiguity: the agent may persist user requests and context to the TaskBoard backend without a clear, user-initiated opt-in, increasing the risk of unintended data storage and action execution.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The English workflow repeats the same ambiguous auto-activation rule, so the issue is not limited to one language section. Because this guidance is intended for AGENTS.md, it is likely to be operationalized directly, causing automatic persistence and workflow execution beyond explicit user intent.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The workflow directs the agent to append final results and task content to persistent cards, but the README does not clearly warn users that their inputs and outputs will be stored on the TaskBoard backend. This undermines informed consent and can lead to inadvertent disclosure or retention of sensitive information.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The English workflow encourages automatic card creation and archival of final summaries without disclosing that conversation-derived content may be persisted externally. In practice, this can cause users' operational details, research outputs, or internal information to be copied into a backend system without adequate notice.

Ssd 3

Medium

Confidence: 95% confidence
Finding: The workflow explicitly instructs the agent to copy full context and final results into persistent task cards. This creates a direct natural-language data retention channel where sensitive user content, internal summaries, or confidential outputs may be stored and later exposed to other systems, users, or board viewers.

Ssd 3

Medium

Confidence: 94% confidence
Finding: The recommended AGENTS.md policy combines broad triggering with automatic persistence of conversation details, making the retention behavior likely to happen silently and routinely. In the context of an agent skill, this is more dangerous because it turns ordinary conversational content into durable backend records without a sensitivity gate or explicit confirmation.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.