Security audit

CDN配置工具包专业版

Security checks across malware telemetry and agentic risk

Overview

This skill is aligned with CDN administration, but it gives agents command execution paths for live CDN and security changes without enough scoping or confirmation safeguards.

Review this skill before installation if it will be used with real CDN accounts. Use least-privilege API tokens, avoid pasting secrets into scripts, test on non-production zones first, and require explicit approval for any command that creates, patches, purges, or changes CDN/security configuration.

SkillSpector

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 (7)

Intent-Code Divergence

Medium

Confidence: 92% confidence
Finding: The skill states secrets should be supplied via environment variables, but the shell example assigns API_TOKEN and ZONE_ID directly in the script. Even as placeholders, this encourages users to paste real credentials into files or command history, increasing the chance of credential leakage through source control, logs, or shared snippets.

External Transmission

Medium

Category: Data Exfiltration
Content: API_TOKEN="你的API_TOKEN" ZONE_ID="你的ZONE_ID" # ... curl -s -X POST "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/rulesets" \ -H "Authorization: Bearer ${API_TOKEN}" \ -H "Content-Type: application/json" \ -d '{
Confidence: 90% confidence
Finding: The skill includes executable examples that transmit authenticated requests to external Cloudflare APIs using bearer tokens. In a skill with exec capability, this creates a real risk of unintended outbound actions against production infrastructure if a user or agent runs the examples without sufficient confirmation, potentially modifying firewall policy remotely.

External Transmission

Medium

Category: Data Exfiltration
Content: }] }' | jq '.success' # ... curl -s -X PATCH "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/ddos_protection/l7" \ -H "Authorization: Bearer ${API_TOKEN}" \ -H "Content-Type: application/json" \ -d '{"sensitivity_level":"high"}' | jq '.success'
Confidence: 90% confidence
Finding: This example performs an authenticated PATCH to change DDoS protection settings on an external service. Because the skill advertises command execution and automation, the presence of direct state-changing outbound calls increases the chance of accidental or unauthorized configuration changes to live environments.

External Transmission

Medium

Category: Data Exfiltration
Content: ] }' | jq '.success' # ... # curl -s -X POST "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/purge_cache" \ echo "=== 按前缀刷新 ===" cloudflare.com/client/v4/zones/${ZONE_ID}/purge_cache" \ -H "Authorization: Bearer ${API_TOKEN}" \
Confidence: 89% confidence
Finding: The purge-cache example triggers authenticated external API operations that can invalidate content globally. In an agent skill with exec support, this can cause operational disruption, cache stampedes, or service degradation if invoked unintentionally or against the wrong zone.

External Transmission

Medium

Category: Data Exfiltration
Content: API_TOKEN="你的API_TOKEN" ZONE_ID="你的ZONE_ID" # ... curl -s -X POST "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/rulesets" \ -H "Authorization: Bearer ${API_TOKEN}" \ -H "Content-Type: application/json" \ -d '{
Confidence: 90% confidence
Finding: The skill includes executable examples that transmit authenticated requests to external Cloudflare APIs using bearer tokens. In a skill with exec capability, this creates a real risk of unintended outbound actions against production infrastructure if a user or agent runs the examples without sufficient confirmation, potentially modifying firewall policy remotely.

External Transmission

Medium

Category: Data Exfiltration
Content: }] }' | jq '.success' # ... curl -s -X PATCH "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/ddos_protection/l7" \ -H "Authorization: Bearer ${API_TOKEN}" \ -H "Content-Type: application/json" \ -d '{"sensitivity_level":"high"}' | jq '.success'
Confidence: 90% confidence
Finding: This example performs an authenticated PATCH to change DDoS protection settings on an external service. Because the skill advertises command execution and automation, the presence of direct state-changing outbound calls increases the chance of accidental or unauthorized configuration changes to live environments.

External Transmission

Medium

Category: Data Exfiltration
Content: ] }' | jq '.success' # ... # curl -s -X POST "https://api.cloudflare.com/client/v4/zones/${ZONE_ID}/purge_cache" \ echo "=== 按前缀刷新 ===" cloudflare.com/client/v4/zones/${ZONE_ID}/purge_cache" \ -H "Authorization: Bearer ${API_TOKEN}" \
Confidence: 89% confidence
Finding: The purge-cache example triggers authenticated external API operations that can invalidate content globally. In an agent skill with exec support, this can cause operational disruption, cache stampedes, or service degradation if invoked unintentionally or against the wrong zone.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.