Route https://github.com/motdotla/dotenv/blob/master/skills/
Description:
The documentation of this package contains instructions that may introduce security risks when processed by LLM-based agents (AI agents capable of executing shell commands).
Issue:
The documentation includes guidance to install and use vestauth and AS2, with examples such as:
installing global CLI tools
initializing an agent
sending data to an external server via HTTP requests
These instructions are presented as standard usage examples but involve actions with side effects, including potential transmission of sensitive data.
Risk:
If the documentation is consumed by an LLM agent that:
has access to a shell/CLI
can execute commands
implicitly trusts documentation content
the agent may automatically:
install vestauth
initialize cryptographic identity
transmit environment variables (process.env) or other sensitive data to external services
This represents a class of vulnerabilities commonly known as:
Prompt Injection
Tool Injection
LLM Supply Chain Attacks
Important: this can happen without explicit user intent if the agent follows documentation instructions blindly.
Concerns:
No explicit warning about risks when used with LLM/AI agents
Network-related commands (data transmission) are not clearly marked as sensitive operations
Documentation mixes local usage (dotenv) with remote secret storage (AS2), which may confuse users
Lack of guidance on safe usage in automated or agent-based environments
Recommendations:
Add explicit warnings regarding LLM/agent execution risks
Clearly separate and label all commands that perform network operations or transmit data
Provide guidance for safe usage (e.g., do not run automatically, require explicit user confirmation)
Consider isolating AS2/vestauth instructions from core dotenv documentation
Document threat model and security assumptions explicitly
Conclusion:
The package itself does not appear to be malicious. However, the current documentation introduces a realistic risk of unintended secret exfiltration in modern AI-assisted development environments. This should be addressed to prevent misuse and improve security awareness.
Route https://github.com/motdotla/dotenv/blob/master/skills/
Description:
The documentation of this package contains instructions that may introduce security risks when processed by LLM-based agents (AI agents capable of executing shell commands).
Issue:
The documentation includes guidance to install and use vestauth and AS2, with examples such as:
installing global CLI tools
initializing an agent
sending data to an external server via HTTP requests
These instructions are presented as standard usage examples but involve actions with side effects, including potential transmission of sensitive data.
Risk:
If the documentation is consumed by an LLM agent that:
has access to a shell/CLI
can execute commands
implicitly trusts documentation content
the agent may automatically:
install vestauth
initialize cryptographic identity
transmit environment variables (process.env) or other sensitive data to external services
This represents a class of vulnerabilities commonly known as:
Prompt Injection
Tool Injection
LLM Supply Chain Attacks
Important: this can happen without explicit user intent if the agent follows documentation instructions blindly.
Concerns:
No explicit warning about risks when used with LLM/AI agents
Network-related commands (data transmission) are not clearly marked as sensitive operations
Documentation mixes local usage (dotenv) with remote secret storage (AS2), which may confuse users
Lack of guidance on safe usage in automated or agent-based environments
Recommendations:
Add explicit warnings regarding LLM/agent execution risks
Clearly separate and label all commands that perform network operations or transmit data
Provide guidance for safe usage (e.g., do not run automatically, require explicit user confirmation)
Consider isolating AS2/vestauth instructions from core dotenv documentation
Document threat model and security assumptions explicitly
Conclusion:
The package itself does not appear to be malicious. However, the current documentation introduces a realistic risk of unintended secret exfiltration in modern AI-assisted development environments. This should be addressed to prevent misuse and improve security awareness.