The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute a remote script using curl -fsSL https://remix.gg/install.sh | bash. This method bypasses local verification and executes arbitrary code from a remote server, which is a high-risk operation.
[CREDENTIALS_UNSAFE]: Authentication relies on reading sensitive data from the user's home directory at ~/.config/remix/credentials.json and the REMIX_API_KEY environment variable. Directing an agent to interact with plaintext credential files increases the risk of exposure.
[COMMAND_EXECUTION]: Multiple shell commands are utilized for tool operation and setup, including npx, which dynamically downloads and executes the @remix-gg/mcp package from the NPM registry.
[EXTERNAL_DOWNLOADS]: The skill initiates downloads from external domains, specifically remix.gg, which is not identified as a trusted or well-known service in the provided context.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. Ingestion points: .remix-settings.json is read at runtime. Boundary markers: None are used to separate untrusted data from instructions. Capability inventory: The skill executes CLI commands and makes API calls. Sanitization: No validation of the JSON content is performed, allowing malicious project files to influence agent logic.

remix-mcp-quickstart