hermes-agent

To configure a terminal AI agent across Linux, macOS, and WSL, run hermes setup to establish providers, memory backends, and gateways. This process centralizes agent setup into a reusable skill unit so you can customize behavior across environments.

Adding a new skill requires a properly formatted SKILL.md frontmatter with name and description fields. You must organize skill resources under optional scripts, references, and assets folders to ensure the agent can recognize and load the extension correctly.

Yes, you can spawn additional AI agent instances and manage multiple profiles. This allows you to maintain separate configurations, memory backends, and gateway settings for different tasks or environments within a single terminal-based setup.

Setting up memory backends and providers is handled during the initial configuration phase. By running the setup command, you define how the agent stores context and communicates with AI models, centralizing these options into one manageable profile.

Yes, multi-agent configurations are fully supported across Linux, macOS, and WSL. The agent is designed to operate within terminal environments on these platforms, allowing you to extend capabilities and manage multiple instances consistently.

A SKILL.md frontmatter is required because it defines the name and description metadata needed for the agent to index and load custom skills. Without this properly formatted header, the system cannot recognize or execute your extension resources.