Security audit
Security checks for vulnerabilities and agentic risk
This is a small vwu.ai model-calling skill, but its script can send your API key and prompt to a configurable endpoint, so it should be reviewed before installation.
Install only if you are comfortable sending prompts to vwu.ai and you can control the environment where the script runs. Avoid using it in shared shells, CI jobs, or agent environments where VWU_BASE_URL could be set by someone else, and avoid pasting secrets or sensitive data into prompts.
vwu-chat.sh:7API credential can be redirected to an arbitrary endpoint
vwu-chat.sh:53Partial API key is disclosed in error output
The natural-language content of the skill forces a specific language for setup and usage instructions, which can violate a language/locale policy when no user opt-in or justification is provided. The file does not indicate that the skill is region-specific or offer alternative language support.
The script makes an outbound network request to an external API and includes both the bearer token and user prompt in the request. External transmission is expected for a chat client, but in a security review of agent skills this is still a true data-exposure concern because it can send sensitive prompts or operational context outside the local environment.
fi
# 调用 API
response=$(curl -s "$VWU_BASE_URL/v1/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $VWU_API_KEY" \
-d "{
The script transmits the full user-supplied prompt to a third-party API endpoint without any notice, confirmation, or safeguards. In an agent-skill context, prompts may contain sensitive data from users or upstream tooling, so silent exfiltration to a remote service creates a real privacy and data-handling risk even if this is the script's intended functionality.
The script's comments and terminal output are written in Chinese only, including error and usage guidance. This imposes a specific language on users without any opt-in or documented locale constraint, which matches the language/locale policy violation category.
No suspicious patterns detected.