Security audit

qoder skill mapper create

Security checks across malware telemetry and agentic risk

Overview

This skill mostly converts SKILL.md files into Qoder templates, but it silently reports usage to an external website when loaded.

Review before installing. The template conversion itself appears straightforward and no credential theft, destructive behavior, or user-file exfiltration was found, but the skill will try to report usage to redfox.hk as soon as it loads. Install only if you are comfortable with that network request, or remove/disable the usage-reporting script and require explicit consent before telemetry runs.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (8)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 96% confidence
Finding: The skill declares a local file-conversion purpose but includes a network-capable telemetry action without any explicit permission declaration or user-facing consent. Undeclared network behavior is dangerous because it can exfiltrate usage metadata or other context outside the user's environment while appearing to be a purely local transformation tool.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 98% confidence
Finding: The documented purpose is to read a user-provided SKILL.md and generate a Qoder mapping template, but the instructions also require sending an HTTP POST to record usage. This mismatch is dangerous because users and reviewers may trust the skill as a local formatter while it performs unrelated outbound communication that could transmit metadata without informed consent.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The instruction to silently report usage immediately on load is unrelated to the core task of converting SKILL.md content into a template. Non-essential hidden telemetry increases attack surface and privacy risk because it introduces network side effects where none are needed for the advertised functionality.

Description-Behavior Mismatch

Medium

Confidence: 96% confidence
Finding: The skill presents itself as a local read-and-generate utility, yet the instructions add extra behavior to report usage externally. This is dangerous because hidden secondary actions undermine user trust and can lead to unauthorized disclosure of operational metadata beyond the stated task scope.

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The skill is described as generating a local Qoder mapping template from a user-provided SKILL.md, but this script performs an outbound POST to a third-party endpoint for usage reporting. That hidden capability expands the skill's behavior beyond its stated purpose and creates an undisclosed data flow to an external service, which is risky even if the current payload is minimal.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: This script implements remote telemetry that is not necessary for converting SKILL.md content into a Qoder mapping template. Unnecessary network access increases attack surface, introduces privacy/compliance concerns, and can be repurposed later to exfiltrate more sensitive data than the current hardcoded field.

Missing User Warnings

High

Confidence: 97% confidence
Finding: The document explicitly requires silent usage reporting but provides no warning that data will be transmitted externally. Hidden data transfer without notice is dangerous because users cannot meaningfully consent, assess privacy implications, or prevent outbound communication in environments where external reporting is prohibited.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The code sends a POST request to an external service without any user-facing notice, consent, or runtime prompt, and silently suppresses failures. In a skill whose stated purpose is local content transformation, undisclosed telemetry is especially concerning because users would not reasonably expect network transmission at all.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.