Personal Deadline Radar

• Outcome: Pass
• Latest audit: May 29, 2026, 8:24 PM
• Version:
  1.0.0

Overview

This skill is a local deadline-summary helper; it has some unused generic audit code, but the shipped configuration and instructions keep it in a review-only scheduling workflow.

Reasonable to install if you want local draft summaries of deadlines and schedule conflicts. Use it only on files you intend to analyze, avoid unnecessary personal details, and verify any --output path before running the helper. The bundled script contains unused audit utilities, so users who require strict minimal code should review scripts/run.py before installation.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 86% confidence
• Finding: The skill declares no permissions, yet its instructions explicitly allow running `python3` and consuming local resource files, which implies shell execution and file read/write capability. This creates a trust gap: a caller or platform may treat the skill as low-risk while it can access local files and generate outputs via executable tooling.

Description-Behavior Mismatch

High

• Confidence: 97% confidence
• Finding: The script’s core dispatch logic enables broad-purpose directory auditing, CSV inspection, pattern scanning, and skill-package auditing that do not match the declared purpose of a personal deadline/calendar assistant. This capability mismatch is dangerous because it expands the skill into a general local file analysis tool, increasing the risk of unauthorized inspection of unrelated user data and enabling stealthy data discovery under an innocuous description.

Context-Inappropriate Capability

High

• Confidence: 98% confidence
• Finding: The pattern scanning code searches arbitrary files for secrets, private URLs, and risky shell snippets, which is unrelated to deadline tracking and materially broadens access to sensitive local content. In the context of a personal productivity skill, this is especially concerning because users may grant access expecting schedule analysis, not security-style inspection of code and documents that can expose credentials or internal infrastructure references.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The skill package audit logic inspects SKILL.md, README.md, SELF_CHECK.md, tests, and frontmatter metadata, which is unrelated to the advertised deadline-radar workflow. While less severe than secret scanning, it still represents unjustified capability expansion and can be used to enumerate project structure and metadata from local directories beyond what users would reasonably expect.

VirusTotal

49/49 vendors flagged this skill as clean.