Dodaios is an open-source agent OS for your desktop: it runs your coding agents, browser and apps by voice or chat, on your own machine. Formerly Personal Jarvis.
Python agent-orchestrationai-agentai-coding-assistantclaude-codecoding-agentcomputer-usedesktop-applocal-firstmcppersonal-jarvisvoice-assistantvoice-coding
import { describe, expect, it } from "vitest";
import {
FAILED_TTL_MS,
MAX_COMMAND_ENTRIES,
pruneCommandActivity,
reduceCommandActivity,
RUNNING_TTL_MS,
SETTLED_TTL_MS,
type CommandActivityEntry,
} from "./commandActivity";
const T0 = 1_000_000;
function proposed(
command: string,
impact?: { level?: string; commands?: string },
) {
return {
tool_name: "run_shell",
args: { command },
risk_tier: "monitor",
...(impact ? { impact } : {}),
};
}
function feed(
events: Array<[string, string, unknown, number?]>,
): CommandActivityEntry[] {
let entries: CommandActivityEntry[] = [];
for (const [name, traceId, payload, ts] of events) {
entries =
reduceCommandActivity(entries, name, traceId, payload, ts ?? T0) ??
entries;
}
return entries;
}
describe("reduceCommandActivity", () => {
it("surfaces a proposed run_shell call with its impact badge", () => {
const entries = feed([
[
"ActionProposed",
"t1",
proposed("rm -rf build", {
level: "destructive",
commands: "rm",
}),
],
]);
expect(entries).toHaveLength(1);
expect(entries[0]).toMatchObject({
id: "t1",
command: "rm -rf build",
level: "destructive",
words: "rm",
status: "running",
});
});
it("ignores every other tool", () => {
const result = reduceCommandActivity(
[],
"ActionProposed",
"t1",
{ tool_name: "open_app", args: { app: "notepad" } },
T0,
);
expect(result).toBeNull();
});
it("pairs Executed with its Proposed via trace id", () => {
const entries = feed([
["ActionProposed", "t1", proposed("ls", { level: "read" })],
[
"ActionExecuted",
"t1",
{ tool_name: "run_shell", success: true, duration_ms: 42 },
T0 + 500,
],
]);
expect(entries[0]).toMatchObject({
status: "done",
durationMs: 42,
settledTs: T0 + 500,
});
});
it("marks a failed execution with its error detail", () => {
const entries = feed([
["ActionProposed", "t1", proposed("definitely-missing")],
[
"ActionExecuted",
"t1",
{ tool_name: "run_shell", success: false, error: "exit 127" },
T0 + 300,
],
]);
expect(entries[0]).toMatchObject({ status: "failed", detail: "exit 127" });
});
it("shows a blacklist denial even without a Proposed", () => {
// A blacklist match raises BEFORE ActionProposed — the denial is the
// only event of that run and must still become a visible card.
const entries = feed([
[
"ActionDenied",
"t9",
{ tool_name: "run_shell", reason: "blacklist: rm -rf /*" },
],
]);
expect(entries[0]).toMatchObject({
status: "blocked",
detail: "blacklist: rm -rf /*",
level: "destructive",
});
});
it("missing impact classifies as unknown, never crashes", () => {
const entries = feed([
["ActionProposed", "t1", proposed("some-binary --x")],
]);
expect(entries[0].level).toBe("unknown");
});
it("dedupes a replayed Proposed for the same trace", () => {
const p = proposed("ls", { level: "read" });
const entries = feed([
["ActionProposed", "t1", p],
["ActionProposed", "t1", p],
]);
expect(entries).toHaveLength(1);
});
it("caps the list and keeps running entries alive", () => {
const events: Array<[string, string, unknown, number?]> = [];
for (let i = 0; i < MAX_COMMAND_ENTRIES + 5; i++) {
events.push(["ActionProposed", `t${i}`, proposed("ls")]);
// Settle every entry except the very first one.
if (i > 0) {
events.push([
"ActionExecuted",
`t${i}`,
{ tool_name: "run_shell", success: true },
]);
}
}
const entries = feed(events);
expect(entries.length).toBeLessThanOrEqual(MAX_COMMAND_ENTRIES);
expect(entries.some((e) => e.id === "t0")).toBe(true); // still running
});
});
describe("pruneCommandActivity", () => {
it("expires done entries after their TTL, failures later", () => {
const entries = feed([
["ActionProposed", "ok", proposed("ls")],
["ActionExecuted", "ok", { tool_name: "run_shell", success: true }, T0],
["ActionProposed", "bad", proposed("rm x")],
[
"ActionExecuted",
"bad",
{ tool_name: "run_shell", success: false, error: "exit 1" },
T0,
],
]);
const afterDoneTtl = pruneCommandActivity(entries, T0 + SETTLED_TTL_MS + 1);
expect(afterDoneTtl?.map((e) => e.id)).toEqual(["bad"]);
const afterFailTtl = pruneCommandActivity(
afterDoneTtl ?? [],
T0 + FAILED_TTL_MS + 1,
);
expect(afterFailTtl).toEqual([]);
});
it("expires a running entry whose Executed never arrived", () => {
const entries = feed([["ActionProposed", "t1", proposed("ls")]]);
expect(pruneCommandActivity(entries, T0 + RUNNING_TTL_MS - 1)).toBeNull();
expect(pruneCommandActivity(entries, T0 + RUNNING_TTL_MS + 1)).toEqual([]);
});
it("returns null when nothing changed", () => {
const entries = feed([["ActionProposed", "t1", proposed("ls")]]);
expect(pruneCommandActivity(entries, T0 + 10)).toBeNull();
});
});