youmind-beehiiv-article

• Outcome: Review
• Latest audit: May 26, 2026, 9:32 AM
• Version:
  1.0.0

Overview

This skill mostly matches a Beehiiv writing and publishing tool, but it also exposes destructive post deletion and an unrelated image-generation path that users should review before installing.

Review this carefully before installing. It needs access to your YouMind API key and can act on the Beehiiv account connected in YouMind, including publishing and deleting posts. Use draft mode by default, avoid invoking delete unless you intend to remove a specific post, protect ~/.youmind/config.yaml as a secret, and do not send confidential drafts or source material unless you are comfortable with YouMind, web-search providers, and Beehiiv receiving that content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Description-Behavior Mismatch

Medium

• Confidence: 96% confidence
• Finding: The CLI exposes listing, fetching, template enumeration, updating, and deletion of Beehiiv posts, which exceeds the skill's declared purpose of writing and publishing articles. In an agent context, this broadens the action surface and can enable unintended content discovery, modification, or removal if the agent or prompting layer invokes the wrong command.

Context-Inappropriate Capability

High

• Confidence: 98% confidence
• Finding: The delete/archive command is not aligned with the advertised skill behavior and introduces an unnecessary destructive capability. In a connected publishing environment, a mistaken or manipulated invocation could remove or archive legitimate newsletter content, causing immediate integrity and availability impact.

Description-Behavior Mismatch

High

• Confidence: 93% confidence
• Finding: This file implements a chat-based image generation capability that is unrelated to the stated Beehiiv article-writing/publishing scope. In an agent skill, hidden or undocumented capabilities are dangerous because they expand what the tool can do beyond user expectations and policy review, enabling unintended remote actions and content generation paths that may be triggered by prompts or orchestration logic.

Context-Inappropriate Capability

High

• Confidence: 95% confidence
• Finding: The skill context is Beehiiv newsletter authoring/publishing, but the code includes agent-mode image generation through /createChat and /listMessages. Context-inappropriate capabilities increase risk because they let the skill invoke broader AI-agent behavior than necessary, which can lead to unreviewed outputs, higher spend, prompt-driven misuse, or accidental execution of side functionality not expected in a publishing tool.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The README instructs users to place a live YouMind API key in a plaintext file under their home directory without warning about credential sensitivity, file permissions, or safer alternatives. If the workstation is shared, backed up insecurely, or compromised by other local processes, the key could be exposed and abused to access the connected YouMind/Beehiiv capabilities.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The skill strongly promotes direct publishing and research via connected services without clearly warning that draft content, source material, and metadata will be sent to external platforms such as YouMind, web-search backends, and Beehiiv. Users may unintentionally disclose unpublished or sensitive content because the transmission boundary is not made prominent at the point of use.

Missing User Warnings

Low

• Confidence: 91% confidence
• Finding: The pipeline instructs saving a working draft to a local path under the user's home directory without any user-facing disclosure or consent gate. While this is not inherently malicious, it can unexpectedly persist sensitive draft content on disk, creating privacy and data-retention risks on shared or managed systems.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The pipeline directs the skill to mine the YouMind knowledge base, use web search, and later publish through external APIs, but it provides no explicit privacy warning about transmitting user content or context to third-party services. In a publishing workflow, source material may contain confidential drafts, internal research, or proprietary newsletter content, so silent transmission increases the chance of unintended data exposure.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The delete action executes immediately with only an ID, without any interactive confirmation, warning banner, or secondary verification. This makes accidental destruction much more likely, especially when the skill may be called by an autonomous or semi-autonomous agent acting on ambiguous user instructions.

VirusTotal

66/66 vendors flagged this skill as clean.