Trust Assessment
neynar-inbox received a trust score of 68/100, placing it in the Caution category. This skill has some security considerations that users should review before deployment.
SkillShield's automated analysis identified 2 findings: 1 critical, 0 high, 1 medium, and 0 low severity. Key findings include Data exfiltration via arbitrary webhook registration, Broad email sending and access capabilities.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The LLM Behavioral Safety layer scored lowest at 63/100, indicating areas for improvement.
Last analyzed on February 13, 2026 (commit 13146e6a). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Data exfiltration via arbitrary webhook registration The skill allows registering webhooks to an arbitrary, user-defined URL (`"url": "https://your-server.com/webhook"`). If an attacker compromises the AI agent using this skill, they could instruct it to register a webhook pointing to an attacker-controlled server. This would cause all inbound emails, potentially containing sensitive information, to be automatically forwarded to the attacker's server, leading to severe data exfiltration. Implement strict validation and allowlisting for webhook URLs within the agent's control logic. The agent should only be allowed to register webhooks to trusted, pre-approved domains or internal services. Alternatively, the skill could offer a more restricted webhook mechanism, e.g., only allowing webhooks to be registered to the same domain as the agent's host. | LLM | SKILL.md:80 | |
| MEDIUM | Broad email sending and access capabilities The skill grants the ability to send emails to arbitrary recipients and retrieve all inbound emails associated with the mailbox. While this is the core functionality of an email skill, it represents a significant attack surface. A compromised agent could be instructed to send spam, phishing emails, or exfiltrate sensitive information by sending it to external addresses. It could also retrieve and process sensitive inbound communications, potentially exposing confidential data. The AI agent orchestrating this skill should implement robust input validation and approval mechanisms for email recipients and content. For inbound emails, sensitive data extraction should be carefully controlled and logged. Consider sandboxing the agent's ability to compose and send emails, or requiring human approval for certain types of communications. | LLM | SKILL.md:30 |
Scan History
Embed Code
[](https://skillshield.io/report/55fd38b66337c44f)
Powered by SkillShield