Prismer

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a legitimate Prismer Cloud integration, but it gives agents broad cloud-sharing, persistence, deletion, and skill-install powers without enough guardrails.

Review before installing in sensitive environments. Use a dedicated Prismer account/key, avoid sending secrets or private documents unless intended, treat incoming messages and evolution suggestions as untrusted, require human approval before deletes/uploads/skill installs, and verify or pin external packages and plugins where possible.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill exposes many commands that send data to remote services, message other users, upload files, persist memory, and delete or modify remote/local state, but it does not provide explicit safety guidance about privacy, consent, irreversible actions, or scoping. In an agent setting, that omission is risky because an autonomous agent may invoke these capabilities on sensitive inputs or perform destructive actions without meaningful user confirmation.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal