An agent capability library | Sami HonkonenIn the last post I described how I set up a local LLM and how I create purpose-built agents:<br>Whenever I want AI help with something specific, I make a new directory under ~/projects/ on my Air and just start working with pi. Once I&rsquo;ve done what I want to do, I tell pi to record the process in an AGENTS.md. From that point on, every time I open pi in that directory it reads the file and is immediately ready to continue.

Continuing on that, I&rsquo;ve started building a general library of capabilities for the agents. Pi reads a global AGENTS.md from ~/.pi/agent/AGENTS.md on startup. Mine is a symlink to a CAPABILITIES.md in a repo called agent-docs. It&rsquo;s a capability index: a list of things the agent can do, each with a short description of when to reach for it and a pointer to a doc that explains how. Here&rsquo;s an excerpt:<br>- **Studio machine** — a powerful always-on Mac you can SSH into.<br>Reach for it when you need more horsepower or a stable host for a<br>background service. Read @/Users/sami/projects/agent-docs/STUDIO.md.

- **exe.dev VMs** — on-demand Linux VMs with a public HTTPS proxy.<br>Reach for it when you need a real Linux box or a public URL.<br>Read @/Users/sami/projects/agent-docs/EXE-DEV.md.

- **Browser automation** — drive a real Chrome from the shell.<br>Reach for it when you need to scrape, fill a form, or visually<br>check a deployed page. Read @/Users/sami/projects/agent-docs/AGENT-BROWSER.md.

The agent doesn&rsquo;t load all specific instructions upfront. It reads the index, decides whether the task matches a capability, and only then reads the relevant doc. The docs themselves are ordinary markdown: what the thing is, when to use it, how to use it. I write and update them almost exclusively with AI. agent-docs is itself a purpose-built agent for maintaining the library.<br>I also include a reference to CAPABILITIES.md in the AGENTS.md of individual coding projects, so project agents can reach for the same tools.<br>The current list has seven entries: the Studio, the local LLM, private git hosting on the Studio, exe.dev VMs, browser automation, a personal MCP server, and this blog. Adding a new one is three steps: write the doc, add a line to the index, commit.<br>The idea is that this compounds. Every time I set something up, I write a doc for it. The agents inherit the capability. Over time the agent should become genuinely useful across a wide range of tasks, because the infrastructure behind it keeps growing.

Get new posts by email.<br>Email address