LazyBoy2/crates/grokboy/src/main.rs

853 lines
30 KiB
Rust
Raw Normal View History

2026-09-13 16:38:32 +00:00
mod team_cli;
2026-09-14 09:08:35 +00:00
mod ui;
2026-09-13 16:38:32 +00:00
use anyhow::{anyhow, Context, Result};
use grokboy_core::{
2026-09-13 16:38:32 +00:00
execute_tool, is_completion_tool, load_or_create, max_rounds_budget, max_rounds_total_budget,
messages_char_len, run_agent, save_session, stream_chat, tool_definitions, truncate_messages,
2026-09-14 09:08:35 +00:00
AgentEvent, AgentVerdict, BoxHub, ChatMessage, Config, InputBroker, Runtime, Session,
2026-09-13 16:38:32 +00:00
ToolContext, AGENT_SYSTEM,
};
use serde_json::json;
use std::io::{self, Write};
use std::path::PathBuf;
use std::process::ExitCode;
const CHAT_SYSTEM: &str = "You are GrokBoy, a concise local coding assistant. Prefer clear, short answers. Traditional Chinese is welcome when the user writes in Chinese.";
#[tokio::main]
async fn main() -> ExitCode {
if let Err(err) = run().await {
eprintln!("error: {err:#}");
2026-09-13 16:38:32 +00:00
return if err.downcast_ref::<CancelledExit>().is_some() {
ExitCode::from(130)
} else {
ExitCode::FAILURE
};
}
ExitCode::SUCCESS
}
2026-09-13 16:38:32 +00:00
#[derive(Debug)]
struct CancelledExit;
impl std::fmt::Display for CancelledExit {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.write_str("agent cancelled")
}
}
impl std::error::Error for CancelledExit {}
fn render_agent_event(event: &AgentEvent) {
2026-09-14 09:08:35 +00:00
ui::render_event(event);
2026-09-13 16:38:32 +00:00
}
fn start_input() -> std::sync::Arc<InputBroker> {
let input = InputBroker::stdin();
let signals = input.clone();
tokio::spawn(async move {
while tokio::signal::ctrl_c().await.is_ok() {
signals.interrupt();
}
});
input
}
async fn execute_turn(
config: &Config,
session: &mut Session,
ctx: &ToolContext,
input: &std::sync::Arc<InputBroker>,
) -> Result<AgentVerdict> {
input.begin();
let runtime = Runtime::for_session(session, input.clone());
runtime.set_event_handler(render_agent_event);
let turn_ctx = ctx.clone().with_runtime(runtime.clone());
let result = run_agent(
config,
&mut session.messages,
&turn_ctx,
max_rounds_budget(),
)
.await;
input.end();
runtime.sync_session(session);
session.last_browser_url = turn_ctx
.last_browser_url_value()
.or(session.last_browser_url.clone());
let verdict =
result.unwrap_or_else(|error| AgentVerdict::Failed(format!("runtime failed: {error:#}")));
session.last_verdict = Some(verdict.kind().into());
session.last_message = Some(verdict.message().into());
session.touch();
save_session(session)?;
Ok(verdict)
}
async fn run() -> Result<()> {
let mut args = std::env::args().skip(1).collect::<Vec<_>>();
if args.is_empty() {
print_help();
return Ok(());
}
let cmd = args.remove(0);
match cmd.as_str() {
2026-09-13 16:38:32 +00:00
"serve" => grokboy_core::team::serve().await,
"agents" => team_cli::agents(&args).await,
"agent" if args.first().map(String::as_str) == Some("--name") => {
team_cli::chat(args.get(1).ok_or_else(|| anyhow!("missing agent name"))?).await
}
"chat" => cmd_chat().await,
"run" => cmd_run(&args).await,
"agent" => cmd_agent(&args).await,
"smoke" => cmd_smoke().await,
2026-09-14 09:08:35 +00:00
"computer" => cmd_computer().await,
"web" => grokboy_core::serve_web(grokboy_core::WebListen::default()).await,
"version" | "-V" | "--version" => {
println!("grokboy {}", env!("CARGO_PKG_VERSION"));
Ok(())
}
"help" | "-h" | "--help" => {
print_help();
Ok(())
}
other => {
eprintln!("unknown command: {other}\n");
print_help();
Err(anyhow!("unknown command"))
}
}
}
fn print_help() {
println!(
"\
2026-09-13 16:38:32 +00:00
GrokBoy — minimal local CLI agent
USAGE:
2026-09-14 09:08:35 +00:00
grokboy serve Start named-agent daemon (Unix socket + HTTP API :8787)
2026-09-13 16:38:32 +00:00
grokboy agents create <name> Create persistent agent in current workspace
grokboy agents list List agent expertise
grokboy agent --name <name> Chat with a persistent agent
grokboy chat Interactive streaming chat (no tools)
grokboy run \"<prompt>\" One-shot agent with tools
grokboy run --session <id> \"...\" Continue a saved session
grokboy agent Interactive multi-turn agent REPL (with tools)
grokboy agent --session <id> Resume an agent session
2026-09-14 09:08:35 +00:00
grokboy computer Start my computer (Docker box) and print the viewer URL
grokboy web Start (or attach to) serve + HTTP API on :8787
Frontend is independent: cd web && npm run dev
grokboy smoke Offline checks (no API key required)
grokboy version
grokboy help
ENV:
GROKBOY_API_KEY API key (or XAI_API_KEY / OPENAI_API_KEY)
GROKBOY_BASE_URL default https://api.x.ai/v1
GROKBOY_MODEL default grok-4.6
2026-09-13 16:38:32 +00:00
GROKBOY_CONTEXT_CHARS approximate context byte budget (default 100000)
GROKBOY_SESSIONS_DIR optional session storage directory
GROKBOY_MAX_ROUNDS Progress interval in model requests (default 12)
2026-09-14 09:08:35 +00:00
GROKBOY_MAX_ROUNDS_TOTAL Model request ceiling per user turn (default 5000)
GROKBOY_PROGRESS 0 = silence live progress on stderr (思考/工具/續跑/結束)
2026-09-14 09:08:35 +00:00
GROKBOY_BROWSER_SURFACE Docker by default; local selects the legacy local browser
GROKBOY_BROWSER_HEADED 1 = visible Chromium in legacy local mode
GROKBOY_WEB_BACKEND_URL Optional AiService-compatible web gateway
GROKBOY_WEB_PROVIDER xai = enable native search on an xAI-compatible proxy
GROKBOY_WEB_HOST web UI bind (default 0.0.0.0)
GROKBOY_WEB_PORT web UI port (default 8787)
GROKBOY_WEB_TOKEN optional bearer token for the web UI
GROKBOY_HANDOFF_AUTO 1 = auto-resume handoff (tests); abort = auto-abort
GROKBOY_CONFIRM_AUTO 1 = auto-approve confirm (tests); abort = auto-deny
(falls back to HANDOFF_AUTO if unset)
2026-09-14 09:08:35 +00:00
Tools: send_message, report_progress, update_plan, request_user_input, report_done, report_blocked, grep, glob, web_fetch, await_command,
2026-09-13 16:38:32 +00:00
exec_command, write_stdin, shell, search_files, list_dir, read_file, edit_file, write_file,
request_user_confirm, browser_navigate, browser_snapshot, browser_read_page,
browser_click, browser_type, browser_press, browser_select, browser_scroll, browser_wait,
browser_tabs, browser_upload, browser_download, browser_eval, browser_handoff
Sessions: ~/.grokboy/sessions/<id>.json
2026-09-14 09:08:35 +00:00
Browser defaults to Docker. Legacy local setup: cd tools/playwright && npm i && npx playwright install chromium
Handoff: agent pauses on login/OTP/captcha → you fix in headed Chromium → Enter
2026-09-13 16:38:32 +00:00
REPL: /exit /quit leave; /session show id; /plan show plan; /stop or Ctrl-C stops a turn
During a task: enter additional instructions to steer the next step.
"
);
}
2026-09-14 09:08:35 +00:00
async fn cmd_computer() -> Result<()> {
let ready = BoxHub::new().ensure_ready().await?;
println!("我的電腦已就緒");
println!("{}", BoxHub::viewer_url());
if let Some(msg) = ready["instruction"].as_str() {
eprintln!("{msg}");
}
Ok(())
}
async fn cmd_chat() -> Result<()> {
let config = Config::from_env().map_err(anyhow::Error::msg)?;
println!(
"GrokBoy chat model={} base={}",
config.model, config.base_url
);
println!("Type a message. Empty line or /exit to quit.\n");
let mut history = vec![ChatMessage::system(CHAT_SYSTEM)];
let stdin = io::stdin();
let mut stdout = io::stdout();
loop {
2026-09-14 09:08:35 +00:00
ui::print_you_prompt();
let mut line = String::new();
if stdin.read_line(&mut line).context("stdin")? == 0 {
println!();
break;
}
let input = line.trim();
if input.is_empty() || input == "/exit" || input == "/quit" {
break;
}
history.push(ChatMessage::user(input));
2026-09-14 09:08:35 +00:00
println!();
print!("Grok ");
stdout.flush().ok();
let reply = stream_chat(&config, &history, |delta| {
print!("{delta}");
let _ = io::stdout().flush();
})
.await?;
println!("\n");
history.push(ChatMessage::assistant(reply));
}
Ok(())
}
async fn cmd_run(args: &[String]) -> Result<()> {
let mut session_id: Option<String> = None;
let mut prompt_parts: Vec<String> = Vec::new();
let mut i = 0;
while i < args.len() {
match args[i].as_str() {
"--session" | "-s" => {
i += 1;
let id = args
.get(i)
.ok_or_else(|| anyhow!("--session requires an id"))?;
session_id = Some(id.clone());
}
"--help" | "-h" => {
println!(
"Usage: grokboy run [--session <id>] \"<prompt>\"\n\
Runs a one-shot tool-using agent. Creates a session under ~/.grokboy/sessions/."
);
return Ok(());
}
other => prompt_parts.push(other.to_string()),
}
i += 1;
}
let prompt = prompt_parts.join(" ");
if prompt.trim().is_empty() {
return Err(anyhow!("usage: grokboy run [--session <id>] \"<prompt>\""));
}
let config = Config::from_env().map_err(anyhow::Error::msg)?;
let cwd = std::env::current_dir().context("cwd")?;
let mut session = load_or_create(session_id.as_deref(), &cwd)?;
// Ensure system prompt is present once at the start.
2026-09-13 16:38:32 +00:00
if session
.messages
.first()
.is_some_and(|m| m.role == grokboy_core::Role::System)
{
session.messages[0] = ChatMessage::system(AGENT_SYSTEM);
} else {
session
.messages
.insert(0, ChatMessage::system(AGENT_SYSTEM));
}
session.cwd = cwd.clone();
2026-09-13 16:38:32 +00:00
if matches!(session.last_verdict.as_deref(), Some("done" | "answer")) {
session.plan.clear();
}
session.push(ChatMessage::user(&prompt));
let tool_ctx = ToolContext::new(session.cwd.clone());
2026-09-13 16:38:32 +00:00
if let Some(url) = &session.last_browser_url {
*tool_ctx.last_browser_url.lock().unwrap() = Some(url.clone());
}
let input = start_input();
let verdict = execute_turn(&config, &mut session, &tool_ctx, &input).await?;
tool_ctx.shutdown().await;
session.last_verdict = Some(verdict.kind().into());
session.last_message = Some(verdict.message().into());
if let Some(url) = tool_ctx.last_browser_url_value() {
session.last_browser_url = Some(url);
}
session.touch();
let path = save_session(&session)?;
print_verdict(&verdict);
2026-09-14 09:08:35 +00:00
ui::print_run_footer(verdict.kind(), &session.id, &path);
2026-09-13 16:38:32 +00:00
if matches!(verdict, AgentVerdict::Cancelled(_)) {
return Err(CancelledExit.into());
}
// Non-zero exit on blocked so scripts can detect fail-closed.
2026-09-14 09:08:35 +00:00
if matches!(verdict, AgentVerdict::Waiting(_)) {
eprintln!(
"提示:已暫停等你。下一句話就是回覆: grokboy run --session {} \"你的回覆\"",
session.id
);
return Ok(());
}
2026-09-13 16:38:32 +00:00
if matches!(
verdict,
AgentVerdict::Blocked(_) | AgentVerdict::BudgetExhausted(_) | AgentVerdict::Failed(_)
) {
eprintln_blocked_recovery_hint_run(&session.id);
2026-09-13 16:38:32 +00:00
return Err(anyhow!("agent {}: {}", verdict.kind(), verdict.message()));
}
Ok(())
}
/// Parse `agent` CLI flags. Returns (session_id, show_help).
fn parse_agent_args(args: &[String]) -> Result<(Option<String>, bool)> {
let mut session_id: Option<String> = None;
let mut i = 0;
while i < args.len() {
match args[i].as_str() {
"--session" | "-s" => {
i += 1;
let id = args
.get(i)
.ok_or_else(|| anyhow!("--session requires an id"))?;
session_id = Some(id.clone());
}
"--help" | "-h" => return Ok((session_id, true)),
other => {
return Err(anyhow!(
"unexpected argument: {other}\nusage: grokboy agent [--session <id>]"
));
}
}
i += 1;
}
Ok((session_id, false))
}
async fn cmd_agent(args: &[String]) -> Result<()> {
let (session_id, show_help) = parse_agent_args(args)?;
if show_help {
println!(
"Usage: grokboy agent [--session <id>]\n\
Interactive multi-turn ReAct agent with tools.\n\
Creates/resumes a session under ~/.grokboy/sessions/.\n\
Commands: /exit /quit leave; /session show id; empty line ignored.\n\
Tip: GROKBOY_BROWSER_HEADED=1 for visible Chromium (handoff)."
);
return Ok(());
}
let config = Config::from_env().map_err(anyhow::Error::msg)?;
let cwd = std::env::current_dir().context("cwd")?;
let created_new = session_id.is_none();
let mut session = load_or_create(session_id.as_deref(), &cwd)?;
2026-09-13 16:38:32 +00:00
if session
.messages
.first()
.is_some_and(|m| m.role == grokboy_core::Role::System)
{
session.messages[0] = ChatMessage::system(AGENT_SYSTEM);
} else {
session
.messages
.insert(0, ChatMessage::system(AGENT_SYSTEM));
}
session.cwd = cwd.clone();
// One ToolContext for the whole REPL so browser URL/state carries across turns.
2026-09-14 09:08:35 +00:00
let mut tool_ctx = ToolContext::new(session.cwd.clone());
tool_ctx.hold_background = false;
2026-09-13 16:38:32 +00:00
if let Some(url) = &session.last_browser_url {
*tool_ctx.last_browser_url.lock().unwrap() = Some(url.clone());
}
// Persist early so the printed session id is on disk.
let path = save_session(&session)?;
2026-09-14 09:08:35 +00:00
ui::print_session_header(
&config.model,
&session.id,
if created_new {
"新 session"
} else {
"續跑"
},
);
2026-09-14 09:08:35 +00:00
eprintln!("{}", path.display());
eprintln!("問你時停住,下一句就是答案。背景工作進行中也可先打字。/exit 離開。");
2026-09-13 16:38:32 +00:00
let input_broker = start_input();
loop {
2026-09-14 09:08:35 +00:00
let live = tool_ctx.has_live_background().await || tool_ctx.has_parked_background();
if live {
eprintln!("背景工作進行中 · 完成後會自動繼續 · 也可以先打字");
}
ui::print_you_prompt();
let incoming = if live {
tokio::select! {
line = input_broker.next() => line,
_ = tool_ctx.wait_background_progress() => {
println!();
eprintln!("背景完成,繼續");
None
}
}
} else {
input_broker.next().await
};
let Some(line) = incoming else {
if live {
let verdict =
execute_turn(&config, &mut session, &tool_ctx, &input_broker).await?;
session.last_verdict = Some(verdict.kind().into());
session.last_message = Some(verdict.message().into());
if let Some(url) = tool_ctx.last_browser_url_value() {
session.last_browser_url = Some(url);
}
session.touch();
let _path = save_session(&session)?;
print_verdict(&verdict);
continue;
}
println!();
break;
2026-09-13 16:38:32 +00:00
};
let input = line.trim();
if input.is_empty() {
continue;
}
if input == "/exit" || input == "/quit" {
break;
}
if input == "/session" {
println!("session {}", session.id);
continue;
}
2026-09-13 16:38:32 +00:00
if input == "/plan" {
println!("{}", serde_json::to_string_pretty(&session.plan)?);
continue;
}
if input == "/stop" {
continue;
}
if matches!(session.last_verdict.as_deref(), Some("done" | "answer")) {
session.plan.clear();
}
session.push(ChatMessage::user(input));
2026-09-13 16:38:32 +00:00
let verdict = execute_turn(&config, &mut session, &tool_ctx, &input_broker).await?;
session.last_verdict = Some(verdict.kind().into());
session.last_message = Some(verdict.message().into());
if let Some(url) = tool_ctx.last_browser_url_value() {
session.last_browser_url = Some(url);
}
session.touch();
2026-09-14 09:08:35 +00:00
let _path = save_session(&session)?;
print_verdict(&verdict);
2026-09-13 16:38:32 +00:00
if matches!(
verdict,
AgentVerdict::Blocked(_) | AgentVerdict::BudgetExhausted(_) | AgentVerdict::Failed(_)
) {
eprintln_blocked_recovery_hint_agent();
}
}
2026-09-13 16:38:32 +00:00
tool_ctx.shutdown().await;
session.touch();
let _ = save_session(&session)?;
Ok(())
}
async fn cmd_smoke() -> Result<()> {
println!("GrokBoy smoke (offline P8 auto-continue)…");
let stamp = uuid_like();
let dir = std::env::temp_dir().join(format!("grokboy-smoke-{stamp}"));
std::fs::create_dir_all(&dir).context("temp dir")?;
let ctx = ToolContext::new(dir.clone());
// write_file
let w = execute_tool(
&ctx,
2026-09-14 09:08:35 +00:00
"external_write_file",
&json!({"path": "note.txt", "content": "smoke ok\n第二行"}).to_string(),
)
.await;
let w: serde_json::Value = serde_json::from_str(&w)?;
2026-09-14 09:08:35 +00:00
assert_ok(&w, "external_write_file")?;
println!(" write_file ok");
// read_file
2026-09-14 09:08:35 +00:00
let r = execute_tool(&ctx, "external_read_file", &json!({"path": "note.txt"}).to_string()).await;
let r: serde_json::Value = serde_json::from_str(&r)?;
2026-09-14 09:08:35 +00:00
assert_ok(&r, "external_read_file")?;
if r["content"].as_str() != Some("smoke ok\n第二行") {
return Err(anyhow!("read_file content mismatch: {r}"));
}
println!(" read_file ok");
// list_dir
2026-09-14 09:08:35 +00:00
let l = execute_tool(&ctx, "external_list_dir", &json!({"path": "."}).to_string()).await;
let l: serde_json::Value = serde_json::from_str(&l)?;
2026-09-14 09:08:35 +00:00
assert_ok(&l, "external_list_dir")?;
let names: Vec<&str> = l["entries"]
.as_array()
.context("entries")?
.iter()
.filter_map(|e| e["name"].as_str())
.collect();
if !names.contains(&"note.txt") {
return Err(anyhow!("list_dir missing note.txt: {l}"));
}
println!(" list_dir ok");
// shell
let s = execute_tool(
&ctx,
2026-09-14 09:08:35 +00:00
"external_shell",
&json!({"command": "printf 'hi'; wc -c < note.txt"}).to_string(),
)
.await;
let s: serde_json::Value = serde_json::from_str(&s)?;
2026-09-14 09:08:35 +00:00
assert_ok(&s, "external_shell")?;
if s["exit_code"] != 0 {
return Err(anyhow!("shell exit_code != 0: {s}"));
}
println!(" shell ok");
// path traversal should fail
let bad = execute_tool(
&ctx,
2026-09-14 09:08:35 +00:00
"external_read_file",
&json!({"path": "../outside.txt"}).to_string(),
)
.await;
let bad: serde_json::Value = serde_json::from_str(&bad)?;
if bad.get("error").is_none() {
return Err(anyhow!("expected path traversal error, got {bad}"));
}
println!(" sandbox ok");
// completion tools
let done = execute_tool(
&ctx,
"report_done",
&json!({"message": "smoke done"}).to_string(),
)
.await;
let done: serde_json::Value = serde_json::from_str(&done)?;
if done["status"] != "done" || done["message"] != "smoke done" {
return Err(anyhow!("report_done mismatch: {done}"));
}
let blocked = execute_tool(
&ctx,
"report_blocked",
&json!({"reason": "smoke blocked"}).to_string(),
)
.await;
let blocked: serde_json::Value = serde_json::from_str(&blocked)?;
if blocked["status"] != "blocked" || blocked["reason"] != "smoke blocked" {
return Err(anyhow!("report_blocked mismatch: {blocked}"));
}
if !is_completion_tool("report_done") || !is_completion_tool("report_blocked") {
return Err(anyhow!("is_completion_tool failed"));
}
println!(" completion ok");
// tool definitions present (6 core + confirm + 6 browser incl. handoff)
let defs = tool_definitions();
let n_tools = defs.as_array().map(|a| a.len()).unwrap_or(0);
2026-09-14 09:08:35 +00:00
if n_tools < 46 {
return Err(anyhow!("expected at least 46 tool defs, got {n_tools}"));
}
let tool_names: Vec<&str> = defs
.as_array()
.unwrap()
.iter()
.filter_map(|t| t["function"]["name"].as_str())
.collect();
if !tool_names.contains(&"browser_handoff") {
return Err(anyhow!("browser_handoff missing from tool defs"));
}
if !tool_names.contains(&"request_user_confirm") {
return Err(anyhow!("request_user_confirm missing from tool defs"));
}
2026-09-14 09:08:35 +00:00
if !tool_names.contains(&"send_message") {
return Err(anyhow!("send_message missing from tool defs"));
}
if !tool_names.contains(&"web_fetch") {
return Err(anyhow!("web_fetch missing from tool defs"));
}
if !tool_names.contains(&"spawn_subagent") {
return Err(anyhow!("spawn_subagent missing from tool defs"));
}
println!(" tool defs ok");
// browser helper protocol (no Chromium required)
let repo = std::env::current_dir().unwrap_or_else(|_| PathBuf::from("."));
match grokboy_core::browser_self_test(&repo) {
Ok(_) => println!(" browser proto ok"),
Err(e) => {
// Soft: smoke still passes if helper missing in weird cwd; warn.
println!(" browser proto skip ({e:#})");
}
}
// Fail-closed install hint when navigating without ensuring chromium — oneshot missing url already covered in helper self-test.
let nav = execute_tool(
&ToolContext::new(repo.clone()),
"browser_navigate",
&json!({"url": ""}).to_string(),
)
.await;
let nav_v: serde_json::Value = serde_json::from_str(&nav)?;
// empty url should error from helper; if helper missing, also error with install hint
if nav_v.get("error").is_none() && nav_v.get("blocked").is_none() {
// empty string might still "succeed" oddly — accept either error or blocked
println!(" browser fail note: {nav_v}");
} else {
println!(" browser fail ok");
}
// handoff protocol (auto-resume; no interactive stdin)
// SAFETY: smoke is single-threaded for this env toggle.
let prev_auto = std::env::var("GROKBOY_HANDOFF_AUTO").ok();
unsafe { std::env::set_var("GROKBOY_HANDOFF_AUTO", "1") };
let handoff = execute_tool(
&ToolContext::new(repo.clone()),
"browser_handoff",
&json!({"reason": "smoke handoff check", "timeout_secs": 3}).to_string(),
)
.await;
let handoff_v: serde_json::Value = serde_json::from_str(&handoff)?;
// Fail-closed without Chromium, or resumed with snapshot if browsers installed.
if handoff_v.get("error").is_none()
&& handoff_v.get("blocked").is_none()
&& handoff_v.get("handoff").and_then(|h| h.as_str()) != Some("resumed")
{
return Err(anyhow!("unexpected handoff smoke result: {handoff_v}"));
}
println!(" handoff auto ok");
// abort path
unsafe { std::env::set_var("GROKBOY_HANDOFF_AUTO", "abort") };
let abort_wait = grokboy_core::wait_for_handoff_resume(1);
if !matches!(abort_wait, grokboy_core::HandoffWait::Aborted(_)) {
return Err(anyhow!("expected HandoffWait::Aborted, got {abort_wait:?}"));
}
println!(" handoff abort ok");
match prev_auto {
Some(v) => unsafe { std::env::set_var("GROKBOY_HANDOFF_AUTO", v) },
None => unsafe { std::env::remove_var("GROKBOY_HANDOFF_AUTO") },
}
// request_user_confirm (auto approve/deny; no interactive stdin)
let prev_confirm = std::env::var("GROKBOY_CONFIRM_AUTO").ok();
unsafe { std::env::set_var("GROKBOY_CONFIRM_AUTO", "1") };
let conf = execute_tool(
&ctx,
"request_user_confirm",
&json!({
"reason": "smoke confirm check",
"prompt": "範例草稿",
"timeout_secs": 2
})
.to_string(),
)
.await;
let conf_v: serde_json::Value = serde_json::from_str(&conf)?;
if conf_v["status"] != "approved" || conf_v["approved"] != true {
return Err(anyhow!("request_user_confirm approve mismatch: {conf_v}"));
}
println!(" confirm auto ok");
unsafe { std::env::set_var("GROKBOY_CONFIRM_AUTO", "abort") };
let deny = execute_tool(
&ctx,
"request_user_confirm",
&json!({"reason": "smoke deny check", "timeout_secs": 2}).to_string(),
)
.await;
let deny_v: serde_json::Value = serde_json::from_str(&deny)?;
if deny_v["status"] != "denied" || deny_v["approved"] != false {
return Err(anyhow!("request_user_confirm deny mismatch: {deny_v}"));
}
let deny_wait = grokboy_core::wait_for_user_confirm(1);
if !matches!(deny_wait, grokboy_core::ConfirmWait::Denied(_)) {
return Err(anyhow!("expected ConfirmWait::Denied, got {deny_wait:?}"));
}
println!(" confirm deny ok");
match prev_confirm {
Some(v) => unsafe { std::env::set_var("GROKBOY_CONFIRM_AUTO", v) },
None => unsafe { std::env::remove_var("GROKBOY_CONFIRM_AUTO") },
}
// context truncation
let mut msgs = vec![
ChatMessage::system(AGENT_SYSTEM),
ChatMessage::user("goal"),
ChatMessage::tool("t1", "X".repeat(50_000)),
ChatMessage::tool("t2", "Y".repeat(50_000)),
ChatMessage::tool("t3", "Z".repeat(50_000)),
ChatMessage::tool("t4", "W".repeat(50_000)),
ChatMessage::tool("t5", "V".repeat(50_000)),
ChatMessage::user("goal again"),
];
let before = messages_char_len(&msgs);
truncate_messages(&mut msgs, 10_000);
let after = messages_char_len(&msgs);
if after >= before {
return Err(anyhow!("truncate did not shrink ({before} -> {after})"));
}
if msgs[0].text() != AGENT_SYSTEM {
return Err(anyhow!("truncate dropped system prompt"));
}
println!(" truncation ok");
// session roundtrip
let mut sess = Session::new(dir.clone());
sess.push(ChatMessage::system(AGENT_SYSTEM));
sess.push(ChatMessage::user("smoke"));
let sess_path = dir.join("session.json");
let data = serde_json::to_vec_pretty(&sess)?;
std::fs::write(&sess_path, data)?;
let loaded: Session = serde_json::from_slice(&std::fs::read(&sess_path)?)?;
if loaded.messages.len() != 2 {
return Err(anyhow!("session roundtrip failed"));
}
println!(" session ok");
// agent CLI parse (offline)
let (sid, help) = parse_agent_args(&[]).expect("empty agent args");
if sid.is_some() || help {
return Err(anyhow!("parse_agent_args([]) unexpected"));
}
let sess_args = vec!["--session".to_string(), "abc-123".to_string()];
let (sid, help) = parse_agent_args(&sess_args)?;
if sid.as_deref() != Some("abc-123") || help {
return Err(anyhow!("parse_agent_args --session failed"));
}
let help_args = vec!["--help".to_string()];
let (_, help) = parse_agent_args(&help_args)?;
if !help {
return Err(anyhow!("parse_agent_args --help failed"));
}
// one-turn session plumbing without API: system + user + fake assistant, save/load
let mut agent_sess = Session::new(dir.clone());
agent_sess.push(ChatMessage::system(AGENT_SYSTEM));
agent_sess.push(ChatMessage::user("第一輪"));
agent_sess.push(ChatMessage::assistant("回覆一"));
agent_sess.push(ChatMessage::user("第二輪"));
let agent_path = dir.join("agent-turn.json");
let data = serde_json::to_vec_pretty(&agent_sess)?;
std::fs::write(&agent_path, data)?;
let loaded_agent: Session = serde_json::from_slice(&std::fs::read(&agent_path)?)?;
if loaded_agent.messages.len() != 4 {
return Err(anyhow!("agent multi-turn session plumbing failed"));
}
if loaded_agent.messages[0].text() != AGENT_SYSTEM {
return Err(anyhow!("agent session missing system prompt"));
}
println!(" agent parse ok");
println!(" agent session ok");
// max rounds env resolver (per chunk default 12; total default 48)
let rounds = max_rounds_budget();
if rounds == 0 {
return Err(anyhow!("max_rounds_budget returned 0"));
}
let total = max_rounds_total_budget();
if total == 0 {
return Err(anyhow!("max_rounds_total_budget returned 0"));
}
println!(" max rounds ok (chunk={rounds}, total={total})");
let _ = std::fs::remove_dir_all(&dir);
if Config::from_env().is_ok() {
println!(" (API key present — skipping live call in smoke; use `run` to exercise)");
} else {
println!(" (no API key — live agent not checked; offline smoke passed)");
}
println!("smoke passed");
Ok(())
}
fn print_verdict(verdict: &AgentVerdict) {
2026-09-14 09:08:35 +00:00
ui::print_verdict(verdict);
}
fn eprintln_blocked_recovery_hint_agent() {
eprintln!(
2026-09-13 16:38:32 +00:00
"提示:session 已保存。額度用完可輸入「繼續」;卡住請補充缺少的資訊;請求失敗可修正設定後重試。"
);
}
fn eprintln_blocked_recovery_hint_run(session_id: &str) {
eprintln!(
2026-09-13 16:38:32 +00:00
"提示:session 已保存。可用 grokboy run --session {session_id} \"繼續\" 接續;若卡住或失敗,請先處理上面的原因。"
);
}
fn assert_ok(v: &serde_json::Value, label: &str) -> Result<()> {
if let Some(err) = v.get("error") {
return Err(anyhow!("{label} failed: {err}"));
}
Ok(())
}
fn uuid_like() -> String {
use std::time::{SystemTime, UNIX_EPOCH};
let n = SystemTime::now()
.duration_since(UNIX_EPOCH)
.map(|d| d.as_nanos())
.unwrap_or(0);
format!("{n}")
}
#[cfg(test)]
mod tests {
use super::parse_agent_args;
#[test]
fn agent_args_empty() {
let (sid, help) = parse_agent_args(&[]).unwrap();
assert!(sid.is_none());
assert!(!help);
}
#[test]
fn agent_args_session() {
let args = vec!["--session".into(), "sess-1".into()];
let (sid, help) = parse_agent_args(&args).unwrap();
assert_eq!(sid.as_deref(), Some("sess-1"));
assert!(!help);
}
#[test]
fn agent_args_help() {
let args = vec!["-h".into()];
let (_, help) = parse_agent_args(&args).unwrap();
assert!(help);
}
#[test]
fn agent_args_rejects_extra() {
let args = vec!["nope".into()];
assert!(parse_agent_args(&args).is_err());
}
}