1package agent
2
3import (
4 "bufio"
5 "context"
6 "fmt"
7 "io"
8 "os"
9 "strings"
10 "time"
11
12 "github.com/patriceckhart/zot/internal/agent/extensions"
13 "github.com/patriceckhart/zot/internal/agent/modes"
14 "github.com/patriceckhart/zot/internal/auth"
15 "github.com/patriceckhart/zot/internal/core"
16 "github.com/patriceckhart/zot/internal/extproto"
17 "github.com/patriceckhart/zot/internal/provider"
18 "github.com/patriceckhart/zot/internal/skills"
19 "github.com/patriceckhart/zot/internal/tui"
20)
21
22// interactiveExtHooks is a tiny adapter that lets the extension
23// manager call back into the Interactive instance built later in
24// runInteractive. The forward-declared *modes.Interactive is filled
25// in immediately after manager construction.
26type interactiveExtHooks struct {
27 ivPtr **modes.Interactive
28}
29
30func (h *interactiveExtHooks) iv() *modes.Interactive {
31 if h == nil || h.ivPtr == nil {
32 return nil
33 }
34 return *h.ivPtr
35}
36
37func (h *interactiveExtHooks) Notify(extName, level, message string) {
38 if iv := h.iv(); iv != nil {
39 iv.Notify(extName, level, message)
40 }
41}
42func (h *interactiveExtHooks) Submit(text string) {
43 if iv := h.iv(); iv != nil {
44 iv.Submit(text)
45 }
46}
47func (h *interactiveExtHooks) Insert(text string) {
48 if iv := h.iv(); iv != nil {
49 iv.Insert(text)
50 }
51}
52func (h *interactiveExtHooks) Display(extName, text string) {
53 if iv := h.iv(); iv != nil {
54 iv.Display(extName, text)
55 }
56}
57
58// extToolAdapter bridges *extensions.Manager to the
59// ExtensionToolSource interface declared in build.go (kept narrow to
60// avoid a build->extensions import cycle). One adapter instance per
61// run; used at every Resolve point so re-built agents pick up the
62// same set of extension tools.
63type extToolAdapter struct {
64 mgr *extensions.Manager
65}
66
67func (a *extToolAdapter) Tools() []ExtensionToolInfo {
68 infos := a.mgr.Tools()
69 out := make([]ExtensionToolInfo, len(infos))
70 for i, t := range infos {
71 out[i] = ExtensionToolInfo{
72 Extension: t.Extension,
73 Name: t.Name,
74 Description: t.Description,
75 Schema: t.Schema,
76 }
77 }
78 return out
79}
80
81func (a *extToolAdapter) NewExtensionTool(info ExtensionToolInfo) core.Tool {
82 return extensions.NewTool(a.mgr, extensions.ToolInfo{
83 Extension: info.Extension,
84 Name: info.Name,
85 Description: info.Description,
86 Schema: info.Schema,
87 })
88}
89
90// fanoutAgentEvent translates a core.AgentEvent into the wire-format
91// EventFromHost and pushes it through the extension manager. Only
92// the events that have a clear extension-facing meaning are
93// forwarded; internal-only ones (text_delta, tool_progress) are
94// dropped to keep the per-extension stream sane.
95func fanoutAgentEvent(mgr *extensions.Manager, ev core.AgentEvent) {
96 if mgr == nil {
97 return
98 }
99 switch e := ev.(type) {
100 case core.EvTurnStart:
101 mgr.EmitEvent(extproto.EventFromHost{Event: "turn_start", Step: e.Step})
102 case core.EvToolCall:
103 mgr.EmitEvent(extproto.EventFromHost{
104 Event: "tool_call", ToolID: e.ID, ToolName: e.Name, ToolArgs: e.Args,
105 })
106 case core.EvAssistantMessage:
107 // Concat the visible text portions of the message; binary
108 // blocks (tool_use, etc.) are skipped because subscribers
109 // usually want a string they can grep / display.
110 var text string
111 for _, c := range e.Message.Content {
112 if tb, ok := c.(provider.TextBlock); ok {
113 text += tb.Text
114 }
115 }
116 mgr.EmitEvent(extproto.EventFromHost{Event: "assistant_message", Text: text})
117 case core.EvTurnEnd:
118 ev := extproto.EventFromHost{Event: "turn_end", Stop: string(e.Stop)}
119 if e.Err != nil {
120 ev.Error = e.Err.Error()
121 }
122 mgr.EmitEvent(ev)
123 }
124}
125
126// Run is the top-level entrypoint for the zot binary.
127func Run(rawArgs []string, version string) error {
128 // Subcommand router: `zot bot ...` is handled separately so the
129 // generic flag parser doesn't reject "bot" as a positional arg.
130 if handled, err := runBotCommand(rawArgs, version); handled {
131 return err
132 }
133 if handled, err := runExtCommand(rawArgs); handled {
134 return err
135 }
136 // `zot rpc` is shorthand for `zot --rpc` so third-party apps can
137 // spawn the binary with a clean argv. Strip the leading 'rpc'
138 // token and let the rest flow through the normal arg parser.
139 if len(rawArgs) > 0 && rawArgs[0] == "rpc" {
140 rawArgs = append([]string{"--rpc"}, rawArgs[1:]...)
141 }
142
143 args, err := ParseArgs(rawArgs)
144 if err != nil {
145 PrintHelp(version)
146 return err
147 }
148 if args.Help {
149 PrintHelp(version)
150 return nil
151 }
152 if args.Version {
153 fmt.Println("zot", version)
154 return nil
155 }
156 // Model catalog: load any cached discovery data before we inspect
157 // the model list (list-models, print/json, interactive).
158 LoadCachedModels()
159
160 if args.ListModels {
161 printModels()
162 return nil
163 }
164
165 ctx := context.Background()
166
167 // Kick an async refresh of the live model catalog. The first run of
168 // zot hits the network; subsequent runs within CacheTTL do nothing.
169 RefreshModelsAsync()
170
171 switch args.Mode {
172 case ModePrint:
173 return runPrintMode(ctx, args, version)
174 case ModeJSON:
175 return runJSONMode(ctx, args, version)
176 case ModeRPC:
177 return runRPCMode(ctx, args, version)
178 default:
179 return runInteractive(ctx, args, version)
180 }
181}
182
183// ---- print / json modes: require credentials, run single-shot ----
184
185func runPrintMode(ctx context.Context, args Args, version string) error {
186 r, err := Resolve(args, true)
187 if err != nil {
188 return err
189 }
190 ag := r.NewAgent()
191 sess, _ := openOrCreateSession(args, r, ag, version)
192 defer sess.Close()
193
194 prompt := args.Prompt
195 if prompt == "" {
196 piped, _ := readAllStdin()
197 prompt = strings.TrimSpace(piped)
198 }
199 if prompt == "" {
200 return fmt.Errorf("print mode requires a prompt (arg or stdin)")
201 }
202
203 start := len(ag.Messages())
204 err = modes.RunPrint(ctx, ag, prompt, nil, os.Stdout)
205 WriteNewTranscript(ag, sess, start)
206 return err
207}
208
209func runJSONMode(ctx context.Context, args Args, version string) error {
210 r, err := Resolve(args, true)
211 if err != nil {
212 return err
213 }
214 ag := r.NewAgent()
215 sess, _ := openOrCreateSession(args, r, ag, version)
216 defer sess.Close()
217
218 prompt := args.Prompt
219 if prompt == "" {
220 piped, _ := readAllStdin()
221 prompt = strings.TrimSpace(piped)
222 }
223 if prompt == "" {
224 return fmt.Errorf("json mode requires a prompt (arg or stdin)")
225 }
226
227 start := len(ag.Messages())
228 err = modes.RunJSON(ctx, ag, prompt, nil, os.Stdout)
229 WriteNewTranscript(ag, sess, start)
230 return err
231}
232
233// ---- interactive mode: opens the TUI even without credentials ----
234
235func runInteractive(ctx context.Context, args Args, version string) error {
236 // Resolve WITHOUT requiring credentials.
237 r, err := Resolve(args, false)
238 if err != nil {
239 return err
240 }
241
242 authStore := AuthStoreFor()
243 mgr := auth.NewManager(authStore)
244 defer mgr.Close()
245
246 // Keep the sandbox pointer stable across agent rebuilds (login / model
247 // switch). The Interactive UI toggles the lock via this pointer, and
248 // rebuilt tool instances must share the same one so the lock sticks.
249 sharedSandbox := r.Sandbox
250
251 // Build the extension manager BEFORE the agent so we can fold
252 // extension-defined tools into the registry. Forward-declare iv so
253 // the host hooks adapter can dereference it after construction.
254 var iv *modes.Interactive
255 extHooks := &interactiveExtHooks{ivPtr: &iv}
256 extMgr := extensions.New(ZotHome(), r.CWD, version, r.Provider, r.Model, extHooks)
257 discoveryErrs := extMgr.Discover(ctx)
258 for _, e := range discoveryErrs {
259 fmt.Fprintln(os.Stderr, "extension load:", e)
260 }
261 // Wait briefly for extensions to flush their initial register_tool
262 // frames before we build the agent's tool registry. Half a second
263 // is plenty for any extension that's actually well-behaved; ones
264 // that don't send a ready frame eat the full grace and proceed.
265 // 3s is the per-extension grace period for the ready frame.
266 // Native binaries are instant; runtimes like `npx tsx` take ~1.5s
267 // from cold cache. The wait is tight only for extensions that
268 // haven't sent ready by then; ones that signalled earlier release
269 // the wait immediately.
270 extMgr.WaitForReady(3 * time.Second)
271 defer extMgr.Stop(2 * time.Second)
272
273 extToolAdapter := &extToolAdapter{mgr: extMgr}
274 r.MergeExtensionTools(extToolAdapter)
275
276 // Capture current args in a closure so BuildAgent can re-resolve
277 // after a successful login (picks up the newly stored credential).
278 wireAgentExt := func(a *core.Agent) *core.Agent {
279 if a == nil {
280 return a
281 }
282 a.BeforeToolExecute = func(call provider.ToolCallBlock) (bool, string) {
283 return extMgr.InterceptToolCall(ctx, call.ID, call.Name, call.Arguments)
284 }
285 a.OnEvent = func(ev core.AgentEvent) { fanoutAgentEvent(extMgr, ev) }
286 return a
287 }
288
289 buildAgent := func() (*core.Agent, string, string, error) {
290 resolved, err := Resolve(args, true)
291 if err != nil {
292 return nil, "", "", err
293 }
294 resolved.UseSandbox(sharedSandbox)
295 resolved.MergeExtensionTools(extToolAdapter)
296 return wireAgentExt(resolved.NewAgent()), resolved.Provider, resolved.Model, nil
297 }
298
299 // Rebuild agent with an explicit provider/model override.
300 buildAgentFor := func(providerOverride, modelOverride string) (*core.Agent, string, string, error) {
301 next := args
302 if providerOverride != "" {
303 next.Provider = providerOverride
304 }
305 if modelOverride != "" {
306 next.Model = modelOverride
307 }
308 resolved, err := Resolve(next, true)
309 if err != nil {
310 return nil, "", "", err
311 }
312 resolved.UseSandbox(sharedSandbox)
313 resolved.MergeExtensionTools(extToolAdapter)
314 return wireAgentExt(resolved.NewAgent()), resolved.Provider, resolved.Model, nil
315 }
316
317 var ag *core.Agent
318 if r.HasCredential() {
319 ag = wireAgentExt(r.NewAgent())
320 }
321
322 // Fire session_start once we know the manager's running.
323 extMgr.EmitEvent(extproto.EventFromHost{Event: "session_start"})
324
325 var sess *core.Session
326 var sessBaselineMsgs int // messages already on disk when current session opened
327 if !args.NoSess && ag != nil {
328 sess, _ = openOrCreateSession(args, r, ag, version)
329 if ag != nil {
330 sessBaselineMsgs = len(ag.Messages())
331 }
332 }
333 defer func() {
334 if sess != nil {
335 sess.Close()
336 }
337 }()
338
339 // loadSession replaces the current session with the one at path and
340 // hands its messages to the agent. Used by the /sessions picker.
341 loadSession := func(path string) error {
342 currentAg := ag // captured
343 if currentAg == nil {
344 return fmt.Errorf("no agent running; log in first")
345 }
346 newSess, msgs, err := core.OpenSession(path)
347 if err != nil {
348 return err
349 }
350 // Flush any unsaved messages to the old session before swapping.
351 if sess != nil {
352 WriteNewTranscript(currentAg, sess, sessBaselineMsgs)
353 _ = sess.Close()
354 }
355 sess = newSess
356 currentAg.SetMessages(msgs)
357 sessBaselineMsgs = len(msgs)
358 return nil
359 }
360
361 term := tui.NewProcTerm()
362
363 // Kick off the async update check so the banner can appear when the
364 // http response eventually arrives (usually <1s on cached DNS). Map
365 // agent.UpdateInfo -> modes.UpdateInfo here to avoid a cyclic import.
366 updateCh := make(chan modes.UpdateInfo, 1)
367 go func() {
368 defer close(updateCh)
369 src := <-CheckForUpdateAsync(ZotHome(), version)
370 updateCh <- modes.UpdateInfo{
371 Current: src.Current,
372 Latest: src.Latest,
373 Available: src.Available,
374 URL: src.URL,
375 }
376 }()
377
378 iv = modes.NewInteractive(modes.InteractiveConfig{
379 Terminal: term,
380 Theme: tui.Dark,
381 Model: r.Model,
382 Provider: r.Provider,
383 AuthMethod: r.AuthMethod,
384 BaseURL: r.BaseURL,
385 Reasoning: r.Reasoning,
386 SystemPrompt: r.SystemPrompt,
387 Tools: r.ToolRegistry,
388 MaxSteps: r.MaxSteps,
389 CWD: r.CWD,
390 ZotHome: ZotHome(),
391 Version: version,
392 UpdateInfoChan: updateCh,
393 Sandbox: sharedSandbox,
394 Agent: ag,
395 InitialInput: args.Prompt,
396 AuthManager: mgr,
397 BuildAgent: buildAgent,
398 BuildAgentFor: buildAgentFor,
399 LoadSession: loadSession,
400 Extensions: extMgr,
401 SkillSnapshot: func() []*skills.Skill {
402 // Re-discover so the picker reflects edits made during
403 // the session. Cheap; SKILL.md files are small.
404 userHome, _ := os.UserHomeDir()
405 list, _ := skills.Discover(ZotHome(), r.CWD, userHome)
406 return list
407 },
408 PersistModel: func(providerName, model string) {
409 // Update config.json so next launch uses the same pick.
410 cfg, _ := LoadConfig()
411 cfg.Provider = providerName
412 cfg.Model = model
413 _ = SaveConfig(cfg)
414 // Update the active session's meta so resume picks this up.
415 if sess != nil {
416 _ = sess.UpdateModel(providerName, model)
417 }
418 },
419 })
420
421 runErr := iv.Run(ctx)
422
423 // Flush final transcript to session (only if we had / ended up with an agent).
424 if finalAg := iv.Agent(); finalAg != nil && sess != nil {
425 WriteNewTranscript(finalAg, sess, sessBaselineMsgs)
426 }
427 return runErr
428}
429
430// openOrCreateSession returns a session for the run. sess may be nil
431// with a nil error if session persistence is disabled.
432func openOrCreateSession(args Args, r Resolved, ag *core.Agent, version string) (*core.Session, error) {
433 if args.NoSess {
434 return nil, nil
435 }
436 // Sweep meta-only files left over from older zot versions (and from
437 // any session that crashed before its first AppendMessage). Cheap;
438 // reads the first few bytes of each file in the cwd's session dir.
439 core.PruneEmptySessions(ZotHome(), args.CWD)
440 var (
441 s *core.Session
442 msgs []provider.Message
443 err error
444 )
445 switch {
446 case args.Session != "":
447 s, msgs, err = core.OpenSession(args.Session)
448 case args.Continue:
449 latest := core.LatestSession(ZotHome(), args.CWD)
450 if latest != "" {
451 s, msgs, err = core.OpenSession(latest)
452 }
453 case args.Resume:
454 picked, perr := pickSession(args.CWD)
455 if perr != nil {
456 return nil, perr
457 }
458 if picked != "" {
459 s, msgs, err = core.OpenSession(picked)
460 }
461 }
462 if err != nil {
463 return nil, err
464 }
465 if s != nil {
466 ag.SetMessages(msgs)
467 return s, nil
468 }
469 return core.NewSession(ZotHome(), args.CWD, r.Provider, r.Model, version)
470}
471
472func pickSession(cwd string) (string, error) {
473 files := core.ListSessions(ZotHome(), cwd)
474 if len(files) == 0 {
475 fmt.Fprintln(os.Stderr, "no sessions for", cwd)
476 return "", nil
477 }
478 for i, f := range files {
479 fmt.Fprintf(os.Stderr, " %2d) %s\n", i+1, f)
480 }
481 fmt.Fprint(os.Stderr, "pick #: ")
482 rd := bufio.NewReader(os.Stdin)
483 line, _ := rd.ReadString('\n')
484 line = strings.TrimSpace(line)
485 var n int
486 if _, err := fmt.Sscanf(line, "%d", &n); err != nil || n < 1 || n > len(files) {
487 return "", fmt.Errorf("invalid selection")
488 }
489 return files[n-1], nil
490}
491
492// WriteNewTranscript appends only messages after index `from` from the
493// agent's transcript to the session.
494func WriteNewTranscript(ag *core.Agent, sess *core.Session, from int) {
495 if sess == nil || ag == nil {
496 return
497 }
498 msgs := ag.Messages()
499 for i := from; i < len(msgs); i++ {
500 _ = sess.AppendMessage(msgs[i])
501 }
502 cum := ag.Cost()
503 _ = sess.AppendUsage(cum, cum)
504}
505
506func readAllStdin() (string, error) {
507 fi, err := os.Stdin.Stat()
508 if err != nil {
509 return "", err
510 }
511 if (fi.Mode() & os.ModeCharDevice) != 0 {
512 return "", nil
513 }
514 b, err := io.ReadAll(os.Stdin)
515 return string(b), err
516}
517
518func printModels() {
519 fmt.Println("provider model id context max-out reasoning source name")
520 for _, m := range provider.Active() {
521 reason := " "
522 if m.Reasoning {
523 reason = "✓"
524 }
525 source := m.Source
526 if source == "" {
527 source = "catalog"
528 }
529 if m.Speculative {
530 source = "speculative"
531 }
532 fmt.Printf("%-10s %-30s %8d %8d %s %-11s %s\n",
533 m.Provider, m.ID, m.ContextWindow, m.MaxOutput, reason, source, m.DisplayName)
534 }
535}