The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill recommends installing the tool using the command curl -sfS https://dotenvx.sh | sh in both the SKILL.md file and the assets/gh_action_example.yaml file. This pattern executes a script from a remote URL directly in the shell without any integrity checks or version pinning, which is a significant security risk if the remote source or the delivery channel is compromised.
[COMMAND_EXECUTION]: The skill provides detailed bash scripts for key rotation that involve executing several high-privilege or sensitive commands, such as gh secret set to update CI secrets and sed to modify configuration files. While these are presented as a tutorial, an agent executing these could inadvertently modify production credentials.
[PROMPT_INJECTION]: The skill documents how to process external data from untrusted .env files which are then used to influence command execution through the dotenvx run command. This represents an indirect prompt injection surface.
Ingestion points: .env and .env.production files (referenced in SKILL.md and assets/env.example).
Boundary markers: None; values from the environment files are directly injected into the execution context of the command.
Capability inventory: Execution of arbitrary subprocesses (node, npm, npm start) with environment variables provided by the tool.
Sanitization: There is no mention of sanitization or validation of the values read from the .env files before they are used in the shell environment.

dotenvx