Competitor Message Differ

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:21 PM UTC
• Version:
  1.0.0

Overview

This skill mainly generates local competitor-messaging comparison drafts, with some unused broader audit code that users should avoid enabling.

Install only if you are comfortable with a local python3 script reading the input you provide and optionally writing a report. Use it for competitor messaging materials, avoid sensitive directories, and do not modify resources/spec.json to enable the unused audit modes unless you intentionally want local file auditing.

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 (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 88% confidence
• Finding: The skill declares no permissions, yet the content explicitly allows use of `python3` with input/output files and references local resources, which implies file read, file write, and shell execution capabilities. This is dangerous because downstream systems or reviewers may trust the undeclared low-risk profile while the skill can in practice access local files and invoke an interpreter, increasing the chance of unintended data access or command execution.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 95% confidence
• Finding: The static finding indicates behavior far beyond the stated purpose of competitor messaging comparison, including directory scanning, CSV/TSV analysis, regex-based secret/high-risk pattern scanning, file validation, and dynamically switching audit modes from `spec.json`. That mismatch is dangerous because it can mask a broader local inspection/auditing tool as a benign analysis skill, causing users to expose unrelated workspace data and enabling overcollection of sensitive information under a misleading description.

Description-Behavior Mismatch

High

• Confidence: 97% confidence
• Finding: The dispatch logic enables multiple modes unrelated to the declared competitor-messaging purpose, including directory, CSV, pattern, and skill-package audits. This scope mismatch is dangerous because it allows the skill to inspect arbitrary local files and repositories under a benign-seeming label, increasing the chance of unauthorized data discovery and deceptive operator use.

Context-Inappropriate Capability

High

• Confidence: 98% confidence
• Finding: The pattern_audit function scans arbitrary files/directories for secrets, private URLs, and dangerous shell patterns, which is unrelated to competitor messaging comparison. In this context, it creates a covert local security scanner that can enumerate sensitive material from user-provided paths, making the misleading skill purpose materially more dangerous.

Context-Inappropriate Capability

High

• Confidence: 96% confidence
• Finding: The skill_audit logic inspects package structure, parses SKILL.md frontmatter, and reports compliance details unrelated to competitor-message differencing. While not directly executing code, it broadens access to arbitrary local project contents under false pretenses and can expose internal repository structure and metadata.

Context-Inappropriate Capability

Medium

• Confidence: 93% confidence
• Finding: The directory and file enumeration logic recursively scans arbitrary paths, lists files, reads markdown content, and summarizes headings. For a competitor messaging tool, this is unjustified data access that can reveal sensitive local project names, content themes, and document structure to downstream outputs.

VirusTotal

66/66 vendors flagged this skill as clean.