From 35719799e62a18850efc52d380fb53bf3baa8022 Mon Sep 17 00:00:00 2001 From: tanaka takashi Date: Tue, 11 Aug 2026 05:17:32 +0900 Subject: Implement P2/P3: Plan-Run flow separation, PlannerService, and Windows test fixes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## P3: Session/Planning Foundation ### PlanAsync / RunAsync flow separation (Design §8.4) - AgentRuntime.ExecuteAsync now follows a two-phase lifecycle: 1. Planning: generate Initial TaskGraph via Ollama structured output 2. Running: execute ready Tasks (dependencies satisfied) in critical-path then creation order - Sandbox Task Resource (BeginTask/EndTask) is applied at RunAsync time, not PlanAsync time (Design constraint) - QueuedTaskWorker transitions to Planning; ExecuteAsync transitions to Running internally ### PlannerService (new: src/Agent.Core/Runtime/PlannerService.cs) - PlanAsync: uses Ollama ChatStructuredAsync with JSON Schema to generate a TaskGraph with per-Task requires_external_effect - ReplanAsync: rebuilds TaskGraph after policy deny, preserving completed tasks (Design §8.4) - Fallback: single-task plan with conservative estimates when structured output fails or returns empty - Pure analysis tasks get RequiresExternalEffect=false with at least 1 Task (Design constraint) ### PlanResult / PlanSchema (new: src/Agent.Core/Runtime/PlanResult.cs) - JSON model for LLM structured output during Planning - JSON Schema with tasks array: description, dependencies (by index), budget_class, PERT estimates, requires_external_effect ### requires_external_effect integration (P2 + P3) - Removed keyword heuristic RequiresExternalEffect() from AgentRuntime - Removed tool-name-counting RequiredToolCalls() from AgentRuntime - ToolUseGuard now receives PlannerTask.RequiresExternalEffect (set by Planner during Planning, not by keyword matching) - requires_external_effect is NOT always true; it is a per-Task attribute determined by the Planner (Design constraint) ### RunAsync / ready Task execution - SelectNextReadyTask: dependencies satisfied, not completed/failed, critical path first then creation order - Each PlannerTask executed via ExecutePlannerTaskAsync (LLM chat-with-tools iteration loop) - TaskGraph.AddUsage records per-Task token usage for CPM tracking ### Replan implementation - Policy deny (InvalidOperationException with "denied by sandbox policy") triggers ReplanAsync, up to max 3 replan attempts - Failed task is marked Failed in TaskGraph before replan - Registry transitions back to Running for replan iteration ## P3-A: AgentSession + TaskGraph ownership ### AgentSession expansion (src/Agent.Core/Sessions/AgentSession.cs) - Added Summary, ExecutionResults, TaskGraphs (ConcurrentDictionary) - TaskGraph is per-MCP-Task (1 agent_run = 1 TaskGraph), NOT shared per-Session (Design constraint) - Added CompactConversation, CompactExecutionResults, SetSummary methods for context summary compaction ### ExecutionResult (new: src/Agent.Core/Sessions/ExecutionResult.cs) - Records tool execution results in session for context summary - Fields: TaskId, ToolName, Success, Output ### SessionStore (src/Agent.Core/Sessions/SessionStore.cs) - Added Sessions property for TaskGraphToolHandler to resolve graphs by TaskId across sessions ### sessionId threading - IInternalToolBackend.BeginTaskAsync signature changed to include sessionId parameter - InternalToolDispatcher, WindowsToolBackend, TaskGraphToolHandler updated to thread sessionId through - task_update schema includes requires_external_effect property - McpHost and Program.cs pass SessionStore to TaskRegistry ## P3: Context Summary (Design §9.3) - Context utilization = latest prompt_eval_count / num_ctx (NOT cumulative InputTokens) (Design constraint) - TryContextSummary: compacts old Conversation/ExecutionResults when utilization exceeds 80% threshold - AgentRuntimeSettings: added NumCtx, ContextSummaryThreshold, ConversationKeepLastN, ExecutionResultsKeepLastN - OllamaResult: added PromptEvalCount field - OllamaClient: passes PromptEvalCount from OllamaChatResponse ## P3: ExecutionResults Session recording - ExecutePlannerTaskAsync records each tool execution result via session.AppendExecutionResult for context summary and audit ## Windows test fixes ### SandboxPolicyException catch (src/Agent.Windows/Tools/WindowsToolBackend.cs) - ExecuteAsync now catches SandboxPolicyException and returns ToolExecutionResult(false, ..., RetryAllowed: false) instead of propagating the exception - Fixes Dir_mkdir_outside_approved_root_is_denied and File_edit_rename_rejects_outside_approved_root tests ### LoopbackExemption cleanup (tests/Agent.Windows.Tests/SecurityPolicyTests.cs) - Recovery_after_crash: added pre-clean and try/finally to ensure LoopbackExemption is always removed, even on assertion failure - Independent_process_crash: added pre-clean to remove residual loopback exemption from prior test failures - Fixes 7 Independent_process_crash scenarios that were failing due to LoopbackExemption residue from Recovery_after_crash ## Test updates (tests/Agent.Core.Tests/PromptAndRuntimeTests.cs) - FakeOllamaClient implementations updated with ChatStructuredAsync and ChatWithToolsAsync overrides to support Plan/Run flow - RecordingOllamaClient: returns single-task plan JSON for ChatStructuredAsync, tracks all call types - ToolCallingOllamaClient: returns plan with requires_external_effect=true for file edit tasks - Test expectations updated for Plan+Run call counts and token accounting (Plan tokens not added to registry usage) ## Documentation updates - Design.md §7.7: requires_external_effect is set by Planner during Planning, not keyword heuristic - Design.md §8.4: Plan/Run phase separation, replan after policy deny - Design.md §9.3: context utilization formula clarified - ImplementationPlan.md §10.3: Plan/Run flow, replan, Context Summary ## Test results (all passing) - Agent.Core.Tests: 61/61 - Agent.Planner.Tests: 10/10 - Agent.Windows.Tests (elevated): 41/41 - Build: 0 errors, 0 warnings --- src/Agent.App/Program.cs | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'src/Agent.App') diff --git a/src/Agent.App/Program.cs b/src/Agent.App/Program.cs index 7839e82..1c29ec3 100644 --- a/src/Agent.App/Program.cs +++ b/src/Agent.App/Program.cs @@ -3,6 +3,7 @@ using Agent.App.Configuration; using Agent.App.Logging; using Agent.Core.Localization; using Agent.Core.Prompts; +using Agent.Core.Sessions; using Agent.Mcp; using Agent.Core.Runtime; using Agent.Windows.Diagnostics; @@ -73,7 +74,8 @@ internal static class Program return 2; } - InternalToolDispatcher toolDispatcher = CreateToolDispatcher(configuration, baseDirectory, logDirectory); + var sessionStore = new SessionStore(); + InternalToolDispatcher toolDispatcher = CreateToolDispatcher(configuration, baseDirectory, logDirectory, sessionStore); var runtimeObserver = new JsonLinesRuntimeObserver(logDirectory); string taskStateDirectory = Path.Combine(baseDirectory, "state", "mcp-tasks"); UserOnlyDirectory.Ensure(taskStateDirectory); @@ -89,7 +91,7 @@ internal static class Program var runtimeSettings = new AgentRuntimeSettings(userPrompt, language, skills, "Windows AppContainer, ACL, Job Object, child-process restriction, and WFP sandbox capabilities.", configuration.Agent.MaxIterations, configuration.Agent.SameFailureLimit, locale.Get("ollama.vision_unsupported")); var visionLimits = new VisionLimits(configuration.Mcp.Images.MaxCount, configuration.Mcp.Images.MaxBytesEach, configuration.Mcp.Images.MaxTotalBytes); await McpHost.RunStdioAsync(locale, OllamaSettingsFactory.Create(configuration.Ollama), runtimeSettings, visionLimits, CancellationToken.None, - toolDispatcher, runtimeObserver, taskStoreSettings, hostObserver).ConfigureAwait(false); + toolDispatcher, runtimeObserver, taskStoreSettings, hostObserver, sessionStore: sessionStore).ConfigureAwait(false); return 0; } if (configuration.Mcp.Transport == "http") @@ -101,7 +103,7 @@ internal static class Program var visionLimits = new VisionLimits(configuration.Mcp.Images.MaxCount, configuration.Mcp.Images.MaxBytesEach, configuration.Mcp.Images.MaxTotalBytes); await McpHost.RunHttpAsync(locale, OllamaSettingsFactory.Create(configuration.Ollama), runtimeSettings, visionLimits, configuration.Mcp.Http.Host, configuration.Mcp.Http.Port, CancellationToken.None, toolDispatcher, runtimeObserver, - taskStoreSettings, hostObserver).ConfigureAwait(false); + taskStoreSettings, hostObserver, sessionStore: sessionStore).ConfigureAwait(false); return 0; } throw new ConfigurationException("Unsupported MCP transport."); @@ -118,7 +120,7 @@ internal static class Program } } - private static InternalToolDispatcher CreateToolDispatcher(AgentConfiguration configuration, string baseDirectory, string logDirectory) + private static InternalToolDispatcher CreateToolDispatcher(AgentConfiguration configuration, string baseDirectory, string logDirectory, SessionStore? sessionStore = null) { string stateDirectory = Path.Combine(baseDirectory, "state"); string journalDirectory = Path.Combine(stateDirectory, "journals"); @@ -142,7 +144,7 @@ internal static class Program configuration.Sandbox.Network.Allow.Select(item => (item.Host, (IReadOnlyList)item.Ports)).ToArray(), journalDirectory, backupDirectory, checked((long)configuration.FileEdit.MemoryBackupLimitMb * 1024 * 1024), configuration.Limits.WholeReadBytes, configuration.Limits.SearchMaxResults, configuration.Limits.TailMaxLines); - var backend = new WindowsToolBackend(settings, audit, new TaskGraphToolHandler(configuration.Planner.MaxTasks)); + var backend = new WindowsToolBackend(settings, audit, new TaskGraphToolHandler(configuration.Planner.MaxTasks, sessionStore)); var allowed = InternalToolCatalog.All.Select(static item => item.Name).ToHashSet(StringComparer.Ordinal); return new InternalToolDispatcher(new DenyByDefaultPolicy(allowed), backend); -- cgit v1.2.3