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 // --ext paths first so they win against installed extensions of
258 // the same name (loadOne's first-write-wins semantics).
259 for _, e := range extMgr.LoadExplicit(ctx, args.Exts) {
260 fmt.Fprintln(os.Stderr, "extension load:", e)
261 }
262 // --no-ext skips the global + project-local discovery scan;
263 // explicit --ext paths above are still honoured so you can run
264 // "only this extension" with --no-ext --ext ./x.
265 if !args.NoExt {
266 for _, e := range extMgr.Discover(ctx) {
267 fmt.Fprintln(os.Stderr, "extension load:", e)
268 }
269 }
270 // Wait briefly for extensions to flush their initial register_tool
271 // frames before we build the agent's tool registry. Half a second
272 // is plenty for any extension that's actually well-behaved; ones
273 // that don't send a ready frame eat the full grace and proceed.
274 // 3s is the per-extension grace period for the ready frame.
275 // Native binaries are instant; runtimes like `npx tsx` take ~1.5s
276 // from cold cache. The wait is tight only for extensions that
277 // haven't sent ready by then; ones that signalled earlier release
278 // the wait immediately.
279 extMgr.WaitForReady(3 * time.Second)
280 defer extMgr.Stop(2 * time.Second)
281
282 extToolAdapter := &extToolAdapter{mgr: extMgr}
283 r.MergeExtensionTools(extToolAdapter)
284
285 // Capture current args in a closure so BuildAgent can re-resolve
286 // after a successful login (picks up the newly stored credential).
287 wireAgentExt := func(a *core.Agent) *core.Agent {
288 if a == nil {
289 return a
290 }
291 a.BeforeToolExecute = func(call provider.ToolCallBlock) (bool, string) {
292 return extMgr.InterceptToolCall(ctx, call.ID, call.Name, call.Arguments)
293 }
294 a.OnEvent = func(ev core.AgentEvent) { fanoutAgentEvent(extMgr, ev) }
295 return a
296 }
297
298 buildAgent := func() (*core.Agent, string, string, error) {
299 resolved, err := Resolve(args, true)
300 if err != nil {
301 return nil, "", "", err
302 }
303 resolved.UseSandbox(sharedSandbox)
304 resolved.MergeExtensionTools(extToolAdapter)
305 return wireAgentExt(resolved.NewAgent()), resolved.Provider, resolved.Model, nil
306 }
307
308 // Rebuild agent with an explicit provider/model override.
309 buildAgentFor := func(providerOverride, modelOverride string) (*core.Agent, string, string, error) {
310 next := args
311 if providerOverride != "" {
312 next.Provider = providerOverride
313 }
314 if modelOverride != "" {
315 next.Model = modelOverride
316 }
317 resolved, err := Resolve(next, true)
318 if err != nil {
319 return nil, "", "", err
320 }
321 resolved.UseSandbox(sharedSandbox)
322 resolved.MergeExtensionTools(extToolAdapter)
323 return wireAgentExt(resolved.NewAgent()), resolved.Provider, resolved.Model, nil
324 }
325
326 var ag *core.Agent
327 if r.HasCredential() {
328 ag = wireAgentExt(r.NewAgent())
329 }
330
331 // Fire session_start once we know the manager's running.
332 extMgr.EmitEvent(extproto.EventFromHost{Event: "session_start"})
333
334 var sess *core.Session
335 var sessBaselineMsgs int // messages already on disk when current session opened
336 if !args.NoSess && ag != nil {
337 sess, _ = openOrCreateSession(args, r, ag, version)
338 if ag != nil {
339 sessBaselineMsgs = len(ag.Messages())
340 }
341 }
342 defer func() {
343 if sess != nil {
344 sess.Close()
345 }
346 }()
347
348 // loadSession replaces the current session with the one at path and
349 // hands its messages to the agent. Used by the /sessions picker.
350 loadSession := func(path string) error {
351 currentAg := ag // captured
352 if currentAg == nil {
353 return fmt.Errorf("no agent running; log in first")
354 }
355 newSess, msgs, err := core.OpenSession(path)
356 if err != nil {
357 return err
358 }
359 // Flush any unsaved messages to the old session before swapping.
360 if sess != nil {
361 WriteNewTranscript(currentAg, sess, sessBaselineMsgs)
362 _ = sess.Close()
363 }
364 sess = newSess
365 currentAg.SetMessages(msgs)
366 sessBaselineMsgs = len(msgs)
367 return nil
368 }
369
370 term := tui.NewProcTerm()
371
372 // Kick off the async update check so the banner can appear when the
373 // http response eventually arrives (usually <1s on cached DNS). Map
374 // agent.UpdateInfo -> modes.UpdateInfo here to avoid a cyclic import.
375 updateCh := make(chan modes.UpdateInfo, 1)
376 go func() {
377 defer close(updateCh)
378 src := <-CheckForUpdateAsync(ZotHome(), version)
379 updateCh <- modes.UpdateInfo{
380 Current: src.Current,
381 Latest: src.Latest,
382 Available: src.Available,
383 URL: src.URL,
384 }
385 }()
386
387 iv = modes.NewInteractive(modes.InteractiveConfig{
388 Terminal: term,
389 Theme: tui.Dark,
390 Model: r.Model,
391 Provider: r.Provider,
392 AuthMethod: r.AuthMethod,
393 BaseURL: r.BaseURL,
394 Reasoning: r.Reasoning,
395 SystemPrompt: r.SystemPrompt,
396 Tools: r.ToolRegistry,
397 MaxSteps: r.MaxSteps,
398 CWD: r.CWD,
399 ZotHome: ZotHome(),
400 Version: version,
401 UpdateInfoChan: updateCh,
402 Sandbox: sharedSandbox,
403 Agent: ag,
404 InitialInput: args.Prompt,
405 AuthManager: mgr,
406 BuildAgent: buildAgent,
407 BuildAgentFor: buildAgentFor,
408 LoadSession: loadSession,
409 Extensions: extMgr,
410 SkillSnapshot: func() []*skills.Skill {
411 // Re-discover so the picker reflects edits made during
412 // the session. Cheap; SKILL.md files are small. Filter
413 // out built-in skills — they're hidden from user-facing
414 // surfaces because they're implementation detail; the
415 // model still sees them through the system-prompt
416 // manifest and the skill tool.
417 userHome, _ := os.UserHomeDir()
418 list, _ := skills.Discover(ZotHome(), r.CWD, userHome)
419 return skills.VisibleSkills(list)
420 },
421 PersistModel: func(providerName, model string) {
422 // Update config.json so next launch uses the same pick.
423 cfg, _ := LoadConfig()
424 cfg.Provider = providerName
425 cfg.Model = model
426 _ = SaveConfig(cfg)
427 // Update the active session's meta so resume picks this up.
428 if sess != nil {
429 _ = sess.UpdateModel(providerName, model)
430 }
431 },
432 })
433
434 runErr := iv.Run(ctx)
435
436 // Flush final transcript to session (only if we had / ended up with an agent).
437 if finalAg := iv.Agent(); finalAg != nil && sess != nil {
438 WriteNewTranscript(finalAg, sess, sessBaselineMsgs)
439 }
440 return runErr
441}
442
443// openOrCreateSession returns a session for the run. sess may be nil
444// with a nil error if session persistence is disabled.
445func openOrCreateSession(args Args, r Resolved, ag *core.Agent, version string) (*core.Session, error) {
446 if args.NoSess {
447 return nil, nil
448 }
449 // Sweep meta-only files left over from older zot versions (and from
450 // any session that crashed before its first AppendMessage). Cheap;
451 // reads the first few bytes of each file in the cwd's session dir.
452 core.PruneEmptySessions(ZotHome(), args.CWD)
453 var (
454 s *core.Session
455 msgs []provider.Message
456 err error
457 )
458 switch {
459 case args.Session != "":
460 s, msgs, err = core.OpenSession(args.Session)
461 case args.Continue:
462 latest := core.LatestSession(ZotHome(), args.CWD)
463 if latest != "" {
464 s, msgs, err = core.OpenSession(latest)
465 }
466 case args.Resume:
467 picked, perr := pickSession(args.CWD)
468 if perr != nil {
469 return nil, perr
470 }
471 if picked != "" {
472 s, msgs, err = core.OpenSession(picked)
473 }
474 }
475 if err != nil {
476 return nil, err
477 }
478 if s != nil {
479 ag.SetMessages(msgs)
480 return s, nil
481 }
482 return core.NewSession(ZotHome(), args.CWD, r.Provider, r.Model, version)
483}
484
485func pickSession(cwd string) (string, error) {
486 files := core.ListSessions(ZotHome(), cwd)
487 if len(files) == 0 {
488 fmt.Fprintln(os.Stderr, "no sessions for", cwd)
489 return "", nil
490 }
491 for i, f := range files {
492 fmt.Fprintf(os.Stderr, " %2d) %s\n", i+1, f)
493 }
494 fmt.Fprint(os.Stderr, "pick #: ")
495 rd := bufio.NewReader(os.Stdin)
496 line, _ := rd.ReadString('\n')
497 line = strings.TrimSpace(line)
498 var n int
499 if _, err := fmt.Sscanf(line, "%d", &n); err != nil || n < 1 || n > len(files) {
500 return "", fmt.Errorf("invalid selection")
501 }
502 return files[n-1], nil
503}
504
505// WriteNewTranscript appends only messages after index `from` from the
506// agent's transcript to the session.
507func WriteNewTranscript(ag *core.Agent, sess *core.Session, from int) {
508 if sess == nil || ag == nil {
509 return
510 }
511 msgs := ag.Messages()
512 for i := from; i < len(msgs); i++ {
513 _ = sess.AppendMessage(msgs[i])
514 }
515 cum := ag.Cost()
516 _ = sess.AppendUsage(cum, cum)
517}
518
519func readAllStdin() (string, error) {
520 fi, err := os.Stdin.Stat()
521 if err != nil {
522 return "", err
523 }
524 if (fi.Mode() & os.ModeCharDevice) != 0 {
525 return "", nil
526 }
527 b, err := io.ReadAll(os.Stdin)
528 return string(b), err
529}
530
531func printModels() {
532 fmt.Println("provider model id context max-out reasoning source name")
533 for _, m := range provider.Active() {
534 reason := " "
535 if m.Reasoning {
536 reason = "✓"
537 }
538 source := m.Source
539 if source == "" {
540 source = "catalog"
541 }
542 if m.Speculative {
543 source = "speculative"
544 }
545 fmt.Printf("%-10s %-30s %8d %8d %s %-11s %s\n",
546 m.Provider, m.ID, m.ContextWindow, m.MaxOutput, reason, source, m.DisplayName)
547 }
548}