There's something pleasingly self-referential about configuring an agent by talking to it.
> Remember that [fact about data formats].
Agent: Noted.
... realise nothing happened ...
> Write to your rules file that when I say "remember", I want you to write to the rules file.
Related Posts
I asked Hermes Agent to change how it set thread titles, and it patched its own source code!
Self modifying code is hard to reason about even in a deterministic environment. I'd already limited Hermes access to the todoist API and web search out of an abundance of caution.
Playing with Hermes Agent and it's the first agentic tool where I feel I need to *justify* my changes. It's updating its notes over time.
W: Split those two jobs, they're completely unrelated.
H: Done. I also updated the skill to note that unrelated topics should be split.
AI agents (OpenClaw, Hermes Agent etc) seem to be increasingly standardising their design.
You have a UI for the markdown files writable by the agent (sometimes just Obsidian) plus a set of CLI tools for performing actions.
I've seen so many new CLIs recently too.