Your first agent | Seshat AI

Give Seshat a real task, watch it work with tools, then teach it your conventions with a skill.

An agent is a model that can act: it reads files, runs commands, searches the web, and keeps going until the task is done. Seshat runs that loop for you, and asks before anything sensitive.

Run a task

Open a project folder in your terminal. Seshat works in the folder where you start it.
```
cd path/to/your/project
```
Start a session.
```
seshat chat
```

Ask for something concrete. For example:

List all TODO comments in this codebase and group them by file.

Watch it work. Seshat shows its reasoning, each tool it calls with the result, and the final answer. Press ctrl+t to see its task list.

Teach it your conventions with a skill

A skill is a Markdown file that is added to the agent’s instructions. Put your team’s rules in .seshat/skills/ in your project:

.seshat/skills/
  git-workflow.md       never commit to main, always open a pull request
  security-rules.md     never log secrets, validate external input

The next session follows them. The official collection, seshat-skills, can be installed from any URL directly from the command line.

Use it from your own code

The same runtime can be embedded in a Go application. The Go SDK page goes further:

import (
    "context"
    "fmt"
    "log"
    "os"

    "github.com/KPO-Tech/seshat/pkg/sdk"
)

cfg := sdk.DefaultClientConfig()
cfg.APIKey = os.Getenv("ANTHROPIC_API_KEY")

client, err := sdk.NewClient(cfg)
if err != nil {
    log.Fatal(err)
}
defer client.Close()

resp, err := client.Ask(context.Background(), "Write a Go function that reverses a string", nil)
if err != nil {
    log.Fatal(err)
}
fmt.Println(resp.Content)

Where to go next

Read What is Seshat? to see how the pieces fit together, and the Architecture for what happens inside the loop.