Ae1
High
- Category
- analysis-evasion
- Confidence
- 100% confidence
- Finding
Referenced artifact was not completely inspected
- Content
md - `references/international-search.md` - International search guide
Security audit
Multi Search Engine 2
Security checks for vulnerabilities and agentic risk
Overview
This is a search-helper skill that documents third-party search URLs and examples, with no local code execution, persistence, or hidden privileged behavior found.
Install only if you are comfortable with search queries being sent to the selected external search engine. Avoid putting secrets, internal project names, personal data, or regulated information into searches, and prefer privacy-oriented engines or stricter safe-search settings for sensitive work.
Vulnerability Patterns
- Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
- Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
- Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
- Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
- Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (9)
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 88% confidence
- Finding
SQP-3 applies to all file types and includes language or locale policy violations. The operational documentation and examples are presented in Chinese only, with no indication that users can choose another language or that the skill is intentionally limited to a Chinese-speaking context.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
The examples instruct use of web_fetch against third-party search engines and omit any warning that user queries will be transmitted to external providers, potentially exposing sensitive prompts, internal project names, or personal data. In an agent skill context, users may copy these examples directly, so the lack of disclosure increases the chance of unintended data leakage to search providers with logging and tracking.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 86% confidence
- Finding
This manifest embeds multiple search engines explicitly labeled with the "cn" region, including CN-specific Bing settings, but provides no natural-language indication that the user can choose or opt into this locale behavior. That can violate language/locale policy expectations because the skill appears to impose a specific regional search context by default rather than offering a documented choice.
- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 91% confidence
- Finding
The documentation promotes sending user queries directly to third-party search engines and includes live request examples, but it does not warn that search terms, IP address, user agent, and related metadata may be disclosed to external providers. In a search-integration skill, this is a real privacy weakness because users may submit sensitive terms assuming the tool is local or privacy-preserving.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 71% confidence
- Finding
The file divides engines into 'Domestic' and 'International' and includes region-specific endpoints such as Bing CN and Google HK, but it does not state how users can choose or opt into a locale/region. This can create an implicit locale preference in skill behavior or examples without explicit user choice.
- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
This section provides concrete examples of sending user queries to external search providers, including potentially sensitive searches such as cached/deleted content lookup and query patterns that may expose user interests or targets, without clearly warning that the queries and metadata will be disclosed to third parties. In a search-integration skill, this is contextually relevant functionality, but the lack of an adjacent privacy notice can mislead users about the privacy implications.
- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The document explicitly shows how to disable DuckDuckGo safe search (
kp=-1) without any warning about exposure to adult, graphic, or otherwise unsafe results. While this is not a code-execution or system-compromise issue, it weakens a safety control and could cause unsafe or policy-violating content retrieval in downstream use.- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 98% confidence
- Finding
The Startpage section documents
nj=0to turn off family filtering without any nearby warning about the consequences of disabling this protective filter. In the context of a general-purpose search skill, documenting filter bypasses without safeguards increases the chance that users or calling agents will retrieve inappropriate or unsafe content unintentionally.- Content
Static analysis
No suspicious patterns detected.