Infinity Poke Recipe Generator

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:26 PM UTC
• Version:
  0.1.2

Overview

This skill is a visible Poke recipe drafting helper that reads templates, writes scaffold files, and gives setup instructions, with no hidden persistence, exfiltration, or destructive behavior found.

Install only if you want a local recipe-scaffolding helper. Use a project-local recipes/poke/<slug> output path, review generated files before publishing, confirm any poke.com lookup is acceptable, and only run generated npx poke login or mcp add commands for accounts and endpoints you trust.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 86% confidence
• Finding: The skill instructs the agent to read templates and generate files, which implies file read/write capabilities, but it does not declare those permissions. Undeclared capabilities weaken review and user consent because the skill can access or modify local content without an explicit permission model, making the behavior less transparent and easier to misuse.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 95% confidence
• Finding: The skill claims it performs a uniqueness check against poke.com/recipes and produces a kitchen-ready, publish-readiness-vetted draft, but the provided content only describes a workflow and scaffolding behavior with placeholders. This mismatch can mislead users into trusting outputs as verified or non-duplicative when no actual external validation or readiness check is implemented, which can cause unsafe publication decisions or policy bypass by false assurance.

Missing User Warnings

Low

• Confidence: 76% confidence
• Finding: The instructions explicitly tell the agent to fetch data from an external website, but there is no user-facing disclosure or consent step for that network access. Even though the target is a public site, silent outbound requests can surprise users, leak usage context, and violate least-astonishment expectations for a local drafting skill.

VirusTotal

66/66 vendors flagged this skill as clean.