Security audit

Web Search (DDG)

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward DuckDuckGo web-search helper with expected network queries and an optional save-to-file feature that should be used carefully.

Install if you are comfortable with search terms being sent to DuckDuckGo. Prefer stdout for quick lookups, and when using --output choose a dedicated safe folder, avoid overwriting important files, and treat saved web results as untrusted external content.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 88% confidence
Finding: The skill documentation explicitly instructs writing search results to arbitrary file paths via --output, but the skill declares no permissions or safeguards around filesystem modification. In an agent setting, this mismatch can lead to unexpected local file creation or overwriting if a user or downstream prompt steers the agent to write into sensitive paths.

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The script accepts an arbitrary --output path and writes search results to that location, creating parent directories as needed. In an agent/skill context, this expands the tool from a simple search utility into a filesystem-writing primitive, which can overwrite or plant files in unintended locations if an upstream agent passes untrusted paths.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The skill description is broad enough to match many generic research, fact-checking, and information requests, which increases the chance of automatic invocation in contexts the user did not intend. Over-broad activation expands attack surface because the agent may perform external network queries or file-output behaviors for routine prompts without clear necessity or user awareness.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The "When to Use This Skill" section lists many broad positive triggers but no constraints, exclusions, or negative examples, making accidental or excessive activation likely. In an autonomous agent, ambiguous triggering can cause unnecessary external searches, retrieval of untrusted content, and potential follow-on actions based on that content.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The documentation normalizes saving results directly to user-specified paths without warning about overwrite risk, path safety, or modification of local files. If an agent follows these examples naively, an attacker could induce writes to sensitive or important files, causing data loss, workspace corruption, or persistence of untrusted content.

Static analysis

No suspicious patterns detected.