merge: integrate oss branch — budget gating, MCP back-channel, events, loopback-only bind

Key features from OSS branch:
- Budget gating: rolling per-provider spend caps with BUDGET_EXCEEDED task state
- Agent MCP back-channel: runners mint tokens; /mcp endpoint resolves them
- Chatbot MCP server at /chatbot/mcp (requires api_token in config)
- Event timeline: unified event stream replacing ad-hoc question/summary flows
- Loopback-only default bind (127.0.0.1:8484); external_bind_allowed=true to expose
- repository_url required on task creation (enforces traceability)
- Auto-checker: spawns verification task after each execution

Conflict resolutions:
- serve.go: keep cfg.Runners.XEnabled() conditional registration from main + agentRegistry from oss
- config.go: keep RunnersConfig from main + BudgetConfig/ExternalBindAllowed from oss
- server.go: handleStaticFiles (base-path rewrite) kept; deduplicated duplicate from both sides
- executor/claude.go: accepted oss deletion (ClaudeRunner replaced by ContainerRunner)
- container_test.go: added noopChannel + AgentChannel parameter to Run call

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

1 files changed, 80 insertions, 34 deletions

diff --git a/internal/executor/local.go b/internal/executor/local.go
index 5d874c6..3f20fe7 100644
--- a/internal/executor/local.go
+++ b/internal/executor/local.go
@@ -37,10 +37,17 @@ func (r *LocalRunner) ExecLogDir(execID string) string {
 	return filepath.Join(r.LogDir, execID)
 }
 
-// Run streams a chat completion to stdout.log. The response is wrapped in
-// stream-json envelopes line-by-line so downstream parsers (summary,
-// changestats) read it the same way they read Claude output.
-func (r *LocalRunner) Run(ctx context.Context, t *task.Task, e *storage.Execution) error {
+// maxLocalToolTurns bounds the tool-use loop so a misbehaving model cannot spin
+// forever calling tools without finishing.
+const maxLocalToolTurns = 12
+
+// Run drives a chat completion against the local endpoint with the agent
+// back-channel tools (ask_user/report_summary/spawn_subtask/record_progress)
+// declared. It loops, feeding tool results back as message history, until the
+// model stops calling tools. Assistant text is written to stdout.log in the
+// same stream-json envelope Claude uses so downstream parsers keep working.
+// If the model calls ask_user, the run stops and returns a *BlockedError.
+func (r *LocalRunner) Run(ctx context.Context, t *task.Task, e *storage.Execution, ch AgentChannel) error {
 	if r.Client == nil {
 		return fmt.Errorf("local runner: no LLM client configured")
 	}
@@ -70,56 +77,95 @@ func (r *LocalRunner) Run(ctx context.Context, t *task.Task, e *storage.Executio
 	if sys := strings.TrimSpace(t.Agent.SystemPromptAppend); sys != "" {
 		messages = append(messages, llm.Message{Role: "system", Content: sys})
 	}
-	messages = append(messages, llm.Message{Role: "user", Content: t.Agent.Instructions})
+	// The agent tools are real here, so guide the model toward them with the
+	// same planning preamble the container runners use (unless skip_planning).
+	messages = append(messages, llm.Message{Role: "user", Content: buildAgentInstructions(t, true)})
 
 	temperature := t.Agent.Temperature
 	if temperature == nil && r.DefaultTemperature > 0 {
 		v := r.DefaultTemperature
 		temperature = &v
 	}
-
-	req := llm.ChatRequest{
-		Model:       t.Agent.Model,
-		Messages:    messages,
-		Temperature: temperature,
-		MaxTokens:   t.Agent.MaxTokens,
-	}
+	tools := agentToolDefs()
 
 	start := time.Now()
-	resp, err := r.Client.ChatStream(ctx, req, func(delta string) {
-		if delta == "" {
-			return
+	var totalIn, totalOut int
+	var lastModel, lastFinish string
+	blocked := false
+
+	for turn := 0; turn < maxLocalToolTurns; turn++ {
+		resp, chatErr := r.Client.Chat(ctx, llm.ChatRequest{
+			Model:       t.Agent.Model,
+			Messages:    messages,
+			Temperature: temperature,
+			MaxTokens:   t.Agent.MaxTokens,
+			Tools:       tools,
+		})
+		if chatErr != nil {
+			writeResultLine(stdout, "error", chatErr.Error(), totalIn, totalOut)
+			return fmt.Errorf("local runner: chat: %w", chatErr)
+		}
+		totalIn += resp.PromptTokens
+		totalOut += resp.OutputTokens
+		lastModel, lastFinish = resp.Model, resp.FinishReason
+
+		if resp.Content != "" {
+			writeAssistantTextLine(stdout, resp.Content)
 		}
-		writeAssistantTextLine(stdout, delta)
-	})
-	if err != nil {
-		writeResultLine(stdout, "error", err.Error(), 0, 0)
-		return fmt.Errorf("local runner: chat: %w", err)
-	}
-	elapsed := time.Since(start)
 
-	// Write one consolidated assistant envelope containing the full response.
-	// extractSummary and ParseChangestatFromOutput operate per-line, so a
-	// single envelope with the full text is what they expect to find.
-	if resp.Content != "" {
-		writeAssistantTextLine(stdout, resp.Content)
+		if len(resp.ToolCalls) == 0 {
+			break
+		}
+
+		// Re-feed: record the assistant's tool-call turn, then each tool result.
+		messages = append(messages, llm.Message{Role: "assistant", Content: resp.Content, ToolCalls: resp.ToolCalls})
+		for _, tc := range resp.ToolCalls {
+			result, didBlock, toolErr := dispatchAgentTool(ctx, ch, tc.Function.Name, tc.Function.Arguments)
+			if toolErr != nil {
+				writeResultLine(stdout, "error", toolErr.Error(), totalIn, totalOut)
+				return fmt.Errorf("local runner: tool %s: %w", tc.Function.Name, toolErr)
+			}
+			if didBlock {
+				blocked = true
+				break
+			}
+			messages = append(messages, llm.Message{
+				Role:       "tool",
+				ToolCallID: tc.ID,
+				Name:       tc.Function.Name,
+				Content:    result,
+			})
+		}
+		if blocked {
+			break
+		}
 	}
-	writeResultLine(stdout, "success", "", resp.PromptTokens, resp.OutputTokens)
+	elapsed := time.Since(start)
 
 	e.CostUSD = 0
-	e.TokensIn = int64(resp.PromptTokens)
-	e.TokensOut = int64(resp.OutputTokens)
+	e.TokensIn = int64(totalIn)
+	e.TokensOut = int64(totalOut)
 
 	if r.Logger != nil {
 		r.Logger.Info("local runner completed",
 			"taskID", t.ID,
-			"model", resp.Model,
-			"tokens_in", resp.PromptTokens,
-			"tokens_out", resp.OutputTokens,
-			"finish_reason", resp.FinishReason,
+			"model", lastModel,
+			"tokens_in", totalIn,
+			"tokens_out", totalOut,
+			"finish_reason", lastFinish,
+			"blocked", blocked,
 			"elapsed_ms", elapsed.Milliseconds(),
 		)
 	}
+
+	// If the model asked the user, stop and block. Local resume re-feeds the
+	// conversation (per-runner sovereign state) — not yet wired, so the session
+	// ID is empty and resume starts fresh for now.
+	if q, isBlocked := channelPendingQuestion(ch); isBlocked {
+		return &BlockedError{QuestionJSON: q}
+	}
+
+	writeResultLine(stdout, "success", "", totalIn, totalOut)
 	return nil
 }