Security audit

huawei-cloud-obs-bucket-create

Security checks across malware telemetry and agentic risk

Overview

This skill is purpose-built for Huawei Cloud OBS bucket creation, but it can create cloud resources with public-read access and run remote installer code without strong confirmation or verification safeguards.

Install only if you are comfortable with an agent helping provision Huawei Cloud OBS resources. Review commands before execution, provide the region explicitly instead of reading local config files, verify any KooCLI installer before running it, and require explicit approval before creating public-read buckets.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Context-Inappropriate Capability

Low

Confidence: 88% confidence
Finding: The skill instructs reading the local ~/.obsutilconfig file to infer the region, which expands the skill's scope from bucket creation guidance into local configuration introspection. Even though it does not explicitly request secrets, accessing local config can expose environment-specific data and establishes an unsafe precedent of using host files as input for operational decisions.

Intent-Code Divergence

Medium

Confidence: 92% confidence
Finding: The document says not to extract AK/SK from configuration files, but later tells the operator to use ~/.obsutilconfig to derive region information. This inconsistency weakens the security boundary around sensitive local configuration and may lead users or agents to treat the config file as generally safe to inspect, increasing the chance of accidental credential exposure.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The trigger phrases include generic terms like 'OBS bucket', 'Huawei Cloud bucket', and 'storage creation', which can cause the skill to activate in contexts broader than bucket creation. Overbroad activation increases the risk that the skill runs in unintended conversations and applies operational instructions or permission-related actions without sufficient task confirmation.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The guide instructs users to fetch a shell script over the network with curl and immediately execute it with bash, which is a classic supply-chain risk pattern. If the hosting location, CDN, DNS, TLS trust chain, or the script itself is compromised, users may run arbitrary code on their systems without reviewing it first.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The script automatically creates buckets and assigns ACLs based on the purpose, including `public-read` for `website` and `assets`, without an explicit confirmation step or prominent warning. In a cloud-storage skill, this is risky because users may unintentionally create publicly accessible buckets, increasing the chance of data exposure or insecure default infrastructure being provisioned at scale.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.