Skill Compiler

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate skill compiler, but it points agents to missing helper scripts and can broadly change skill runtime files, so it needs review before installation.

Install only after verifying the missing helper scripts from a trusted source. Run compilation only on an intended skills directory, review generated file diffs before relying on them, and avoid using the generic execute-style triggers or publish command unless the action is explicit and scoped.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The trigger phrase `execute <skill>` is generic and likely to collide with ordinary user instructions, increasing the chance of accidental skill invocation. In this compiler skill, that ambiguity is more concerning because the skill performs filesystem-oriented build actions and may compile or traverse directories when a user merely intended a natural-language command.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The skill advertises very broad trigger phrases (`exe <skill>` and `execute <skill>`) without defining boundaries, exclusions, or disambiguation rules. In an agent environment, generic activation words can cause accidental invocation or prompt collisions with unrelated user instructions, increasing the chance that this operational skill runs in contexts where the user did not clearly intend it.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal