Phoenix Shield

Security checks across malware telemetry and agentic risk

Overview

Phoenix Shield matches its backup-and-rollback purpose, but it gives agents broad production system update, backup, monitoring, and rollback authority without enough safety scoping.

Install only if you are prepared to supervise it like a production administration tool. Verify the actual phoenix-shield CLI source before use, review every command passed to --command or hook flags, test in staging first, limit privileges and target hosts, avoid unattended auto-rollback until proven safe, and protect backups because they may contain secrets and sensitive operational data.

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 (3)

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The skill description is framed broadly enough to be invoked for many common update and deployment tasks, which increases the chance an agent will apply it in high-risk production contexts without sufficient human review. Because the skill performs backups, deployments, and rollback-oriented operations, overbroad triggering can lead to unnecessary privileged system changes and expanded attack surface.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The rollback section describes automatic recovery paths including package downgrade and full system restore, but it does not prominently warn that these actions can overwrite system state, remove recent changes, or cause service interruption. In an automated agent context, that omission is dangerous because users may not realize the skill can make destructive changes during failure handling.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The backup description states that snapshots include configuration files, database dumps, process lists, network connections, and health metrics, all of which may contain secrets, credentials, topology, and operationally sensitive information. Without a clear privacy and data-handling warning, users or agents may collect and retain sensitive data more broadly than intended.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal