SAAS Orchestrator

• Outcome: Review
• Latest audit: May 28, 2026 at 12:26 PM UTC
• Version:
  1.0.0

Overview

This skill is mostly a SaaS planning prompt generator, but it can steer agents toward payments, deployments, and public launches without clear approval controls.

Use this skill for planning unless you deliberately want an agent to act. Review generated subagent prompts before running them, treat the factory status as mock unless connected to real data, and require explicit approval before deployment, payment setup, account creation, or public launch actions. Use sandbox or least-privilege credentials.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 84% confidence
• Finding: The skill advertises orchestration behavior and references scripts that can create or save outputs, yet it declares no permissions despite detected file-write capability. In an agent environment, undeclared write access breaks least-privilege expectations and can allow unreviewed modification of workspace files, logs, prompts, or generated artifacts.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 92% confidence
• Finding: The skill's stated purpose materially overclaims what it actually does, presenting static templates and mock tracking as if it performs real orchestration, monitoring, and revenue management. This can mislead users or higher-level agents into trusting fabricated state, making operational decisions on false data and masking the fact that no real coordination or validation is happening.

VirusTotal

61/61 vendors flagged this skill as clean.