ClawHub
Back to skill

Security audit

rust-review

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Rust code-review skill with some broad triggers and optional command guidance that users should run deliberately.

Install this only if you want an opinionated Rust review assistant. Before letting it run commands, especially cargo install cargo-mutants, cargo mutants --workspace, cargo audit, cargo outdated, or cargo deny check, confirm the project scope and approve the tool/network/resource use explicitly.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Description-Behavior Mismatch

High
Confidence
97% confidence
Finding
This module materially departs from the declared skill scope of Rust security review and instead instructs the agent to perform broad test-quality auditing. In an agentic system, scope drift is dangerous because it can cause the model to spend review budget on irrelevant checks, omit unsafe/ownership/dependency risks it was supposed to assess, and produce misleading outputs under the trusted label of a security skill.

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
The module directs installation and execution of external mutation-testing tooling (`cargo install cargo-mutants` and `cargo mutants --workspace`), which expands the skill from static review into environment-modifying and potentially expensive command execution. In the context of a review skill, this creates unnecessary operational and supply-chain exposure, and could trigger unintended network access, package installation, or CI resource consumption.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger list contains broad single-word activators like "rust," "ownership," "concurrency," "unsafe," "traits," and "cargo," which are likely to match ordinary discussion rather than clear user intent to invoke the skill. This can cause the skill to auto-activate in unrelated contexts, increasing the chance of unnecessary prompt injection exposure, workflow interference, or accidental execution of review behaviors.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The natural-language trigger "reviewing Rust code or before merging Rust changes" is ambiguous and lacks precise activation boundaries. Because it describes a broad situation instead of an explicit invocation pattern, the skill may be invoked opportunistically during ordinary development conversations, which expands the attack surface for unintended activation and instruction interference.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.