Keenable Web Search

• Outcome: Pass
• Latest audit: Jun 27, 2026 · 8:08 PM UTC
• Version: 1.1.2

Overview

This is a straightforward web-search skill that sends search queries to Keenable's public endpoint, with no evidence of hidden persistence, credential access, local file access, or destructive behavior.

Install this only if you are comfortable sending search terms to Keenable's public search endpoint. Do not include secrets, credentials, private customer data, or confidential business material in queries, and consider using explicit Keenable-specific invocation if your agent auto-routes broad search phrases.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 95% confidence
• Finding: The skill explicitly performs outbound web requests via a bundled Python script, but the skill metadata does not declare network permissions. This creates a transparency and governance gap: users or host platforms may invoke the skill without realizing queries are sent to a third-party endpoint, which can expose sensitive prompts or research topics.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The trigger list includes broad, common phrases such as 'research', 'news', 'look it up', and 'web search', which are likely to match ordinary user language and cause the skill to activate unexpectedly. Because this skill performs networked searches, accidental activation can send user queries externally without deliberate intent or adequate consent.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: Although the skill says it searches the live web, it does not present a prominent user-facing privacy warning at the point of use that queries are transmitted to a third-party public endpoint. In practice, users may include sensitive, proprietary, or regulated information in natural-language queries, causing unintended external disclosure.

Shadow Command Trigger

Medium

• Category: Trigger Abuse
• Confidence: 83% confidence
• Finding: The trigger phrase 'search online' is close to a common built-in capability name and can shadow or intercept normal assistant behavior. In environments with automatic routing, this can redirect routine user requests into this third-party networked skill, changing execution semantics and increasing privacy risk.

Shadow Command Trigger

Medium

• Category: Trigger Abuse
• Confidence: 80% confidence
• Finding: The trigger phrase 'find sources' is generic and overlaps with common assistant intent parsing around finding information. This can cause unintended skill invocation and external transmission of user requests to the search service when the user may have expected an internal or built-in capability.

Static analysis

No suspicious patterns detected.