SendGrid Skills

• Outcome: Pass
• Latest audit: Jul 14, 2026 · 4:57 PM UTC
• Version: 1.2.3

Overview

This is a coherent SendGrid integration skill for sending and receiving email, with expected network and API-key use but privacy-sensitive workflows users should handle carefully.

Install only if you intend to use SendGrid. Use a narrowly scoped SendGrid API key, verify sender and recipient details before sending, avoid sending secrets or regulated data unless appropriate, and treat inbound email bodies and attachments as untrusted sensitive data with logging, storage, malware scanning, and retention controls.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (8)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 90% confidence
• Finding: The skill declares executable shell-related capabilities through required binaries and operational notes, but does not expose an explicit permissions model to constrain or communicate those actions. In an agent setting, this can lead to unexpected command execution against local files or network targets, especially since the notes explicitly mention user-provided file paths and endpoints.

Missing User Warnings

Medium

• Confidence: 85% confidence
• Finding: The README promotes workflows for sending transactional email and receiving inbound email, including parsed email content, but does not warn that these flows may process personal data, secrets, message content, or attachments. In an AI-agent skill context, omission of privacy and data-handling cautions increases the chance that users will implement collection, storage, or forwarding of sensitive email data without appropriate consent, minimization, retention, or access controls.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The example prompt explicitly suggests parsing incoming email attachments and storing them in S3 without any warning that attachments can contain highly sensitive documents, malware, secrets, or regulated data. Because this is guidance for an agent skill, users may reproduce the pattern directly and create unsafe ingestion/storage pipelines without scanning, classification, encryption, or restricted access.

Vague Triggers

Medium

• Confidence: 82% confidence
• Finding: The trigger phrases are broad and overlap with common email-related requests such as 'send email' and 'receive email', which increases the chance this skill is invoked in contexts where a narrower or safer skill would be more appropriate. Over-broad routing can expose SendGrid-related shell and network behaviors unintentionally, causing confused-deputy style actions or unnecessary access to sensitive email workflows.

Missing User Warnings

Medium

• Confidence: 83% confidence
• Finding: The README instructs users to parse inbound webhook payloads containing full email bodies, headers, envelope data, and attachments, but it does not warn that these fields may contain sensitive personal, credential, or business data. In a debugging context, users may capture, print, or persist this content to logs or terminals, creating avoidable data exposure risks.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The example workflow suggests testing a live webhook and piping or capturing webhook payload data without warning that the resulting output or logs may expose private email contents and attachments. Because this is an email-processing skill, the context makes the omission more dangerous: users are likely to handle real inbound messages that may include sensitive or regulated information.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger list is broad enough to match generic phrases like 'send email' and 'email notification', which can cause the skill to activate in contexts where the user did not specifically intend to use SendGrid. Because this skill can transmit data externally, unintended invocation increases the risk of accidental outbound disclosure of message content, recipients, or attachments.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill explains how to send emails externally but does not clearly warn that using it will transmit content, recipient addresses, and possible attachments to third parties via SendGrid. In an agent setting, that omission can lead to users or upstream systems invoking the skill without understanding the privacy and data-exposure implications.

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.