Alibabacloud Sls Data Agent

• Outcome: Pass
• Latest audit: Jun 15, 2026 at 2:41 AM UTC
• Version:
  0.0.1

Overview

This skill is a disclosed Alibaba Cloud SLS DataAgent connector, but users should understand that questions and SLS context are sent to Alibaba Cloud using their configured cloud credentials.

Install this only if you intend to use Alibaba Cloud SLS DataAgent. Use least-privilege Alibaba Cloud credentials limited to the needed DataAgent actions and SLS scope, confirm the project/logstore before each analysis, and do not include secrets, regulated data, or unrelated local/private content in questions sent through the skill.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The description emphasizes autonomous analysis but does not prominently warn that natural-language queries and session metadata are transmitted to a remote Alibaba Cloud API over SSE/HTTP and are traceable via a session-specific user-agent. In a data-analysis skill, prompts often contain sensitive operational details, so lack of clear disclosure can lead to unintended exfiltration of log content, identifiers, or investigative context to a third party.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The trigger text is broad enough to match generic requests like 'data analysis' or 'charts', which can cause the skill to activate outside a clearly Alibaba Cloud SLS-specific context. Because the skill is networked and autonomous, over-broad routing raises the chance that unrelated or sensitive user requests are forwarded to the external DataAgent without the user intentionally choosing this integration.

Missing User Warnings

Medium

• Confidence: 83% confidence
• Finding: The skill sends the user's natural-language question, plus project/logstore/skill context, to a remote Alibaba Cloud endpoint without any in-script disclosure, consent gate, or data-sensitivity check. In an agent setting, users may unknowingly submit secrets, proprietary log contents, incident details, or regulated data to an external service, creating privacy, compliance, and data-handling risk.

Missing User Warnings

Medium

• Confidence: 80% confidence
• Finding: The code automatically loads Alibaba Cloud credentials from the default credential chain and uses them to access remote APIs without surfacing that behavior to the user. In a shared agent/runtime environment, this can cause actions to be executed under unexpectedly privileged cloud identities, increasing the chance of unauthorized access, billing impact, or cross-project data exposure.

VirusTotal

65/65 vendors flagged this skill as clean.