Security audit

ucloud cli

Security checks for vulnerabilities and agentic risk

Overview

This UCloud CLI skill is coherent, but it needs review because it can make real cloud changes from broad deployment requests and tells the agent to show login credentials in chat.

Install only if you want an agent to perform real UCloud account operations. Before use, confirm the intended profile, project, region, billing impact, and whether the request should really target UCloud. Avoid letting the agent print passwords in chat; prefer key-based access, a secret manager, or one-time credential handling, and rotate any credential that appears in a transcript.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: The password-generation snippet is outside the core purpose of a UCloud resource-management skill and introduces an unnecessary secret-generation capability. In a skill that already performs real cloud operations, this can encourage ad hoc credential creation without lifecycle controls, auditing, or service-specific password policy validation, expanding the skill’s operational scope in a risky way.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger description is overly broad, covering generic deployment and go-live scenarios rather than narrowly defined UCloud CLI tasks. This can cause the skill to activate in ambiguous contexts and perform or propose real cloud actions when the user did not clearly intend UCloud-specific operations, increasing the risk of unintended infrastructure changes.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: Treating all deployment-like requests as a strong trigger without constraints or negative examples creates ambiguous routing for a high-impact skill capable of real infrastructure changes. In practice, this can misclassify generic app deployment discussions as authorization to manipulate cloud resources, increasing the chance of unintended execution paths.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The activation guidance is overly broad: generic phrases like 'deploy my web service' or 'put this site online' can match many routine user requests and cause the skill to initiate real cloud deployment workflows without sufficiently clear boundaries. In a skill that performs actual UCloud resource operations, ambiguous triggering increases the chance of unintended infrastructure creation, cost incurrence, or execution of higher-risk actions from underspecified prompts.

Missing User Warnings

High

Confidence: 95% confidence
Finding: The skill instructs the agent to display sensitive post-deployment details including login credentials, internal/external IPs, resource IDs, and network/security-group information in a summary without warning or minimization. Natural-language summaries are often logged, echoed to users, or exposed in chat history, so this materially increases the risk of credential leakage and infrastructure reconnaissance.

Ssd 3

High

Confidence: 98% confidence
Finding: Explicitly directing the agent to display newly created resource login credentials in conversational output is a direct secret-disclosure flaw. Credentials exposed in model output can be retained in transcripts, observability tooling, or downstream integrations, enabling unauthorized access to newly provisioned systems.

Static analysis

No suspicious patterns detected.