Implement Phase 1: tool-calling loop, conversation history, CI/CD, docs, resilience, observability by Telli · Pull Request #3 · clawdotnet/SharpClawCode

Telli · 2026-04-11T19:02:24Z

Summary

Closes all 7 Phase 1 gaps from the PRD, transforming SharpClaw Code from a streaming wrapper into a full coding agent runtime with tool-calling, conversation history, and production infrastructure.

P0 — Tool-Calling Loop + Conversation History

Protocol models: ChatMessage, ContentBlock, ContentBlockKind, ProviderToolDefinition for structured conversation history
Provider extensions: ProviderRequest now carries Messages, Tools, MaxTokens; ProviderEvent carries tool-use metadata
Anthropic adapter: Extracts tool_use content blocks from streaming, AnthropicMessageBuilder maps messages/tools to SDK types
OpenAI adapter: Extracts FunctionCallContent from MEAI streaming, OpenAiMessageBuilder maps to MEAI types
ToolCallDispatcher: Bridges tool-use provider events to IToolExecutor with permission enforcement and event publishing
Tool-calling loop: ProviderBackedAgentKernel now implements multi-iteration loop (stream → detect tool-use → execute → feed back → repeat, up to 25 iterations)
Conversation history: ConversationHistoryAssembler builds ChatMessage[] from session events; ContextWindowManager applies token-aware truncation (100K budget)

P1 — Infrastructure

CI/CD: GitHub Actions workflows for build+test (3-platform matrix) and NuGet release on version tags
NuGet metadata: Package descriptions, authors, license, repository URL on all 18 publishable projects
Getting started guide: 15-minute zero-to-running guide with code samples
Agent Framework integration guide: Positioning doc showing SharpClaw as production layer on top of Microsoft Agent Framework
Example projects: Minimal console agent, ASP.NET Core web API agent, custom tool agent

P2 — Production Hardening

Provider resilience: ResilientProviderDecorator with exponential backoff + jitter retry, 429/Retry-After handling, circuit breaker (open/half-open/closed)
Observability: SharpClawActivitySource + TurnActivityScope + ProviderActivityScope (OpenTelemetry spans), SharpClawMeterSource (counters + histograms for tokens, duration, tool invocations), NdjsonTraceFileSink for offline trace analysis

Stats

78 files changed (+3,920 lines)
192 tests passing (152 unit + 25 integration + 15 parity)
0 errors, 0 test failures

Test plan

dotnet build SharpClawCode.sln — 0 errors
dotnet test SharpClawCode.sln — 192/192 pass
Tool-calling parity test (tool_call_roundtrip) verifies end-to-end loop
Resilience tests verify retry, circuit breaker, and non-transient failure behavior
Session storage, state machine, and context window tests cover new runtime code
Example projects build independently
Manual verification of CI workflow on push (will trigger on PR creation)

🤖 Generated with Claude Code

…ion history Introduces ContentBlock (with ContentBlockKind enum) and ChatMessage sealed records, registers them in ProtocolJsonContext, and adds three roundtrip serialization tests covering text, tool-use, and tool-result blocks. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…vent with tool-use fields Add ProviderToolDefinition to Protocol (decoupled from Tools), extend ProviderRequest with Messages/Tools/MaxTokens, extend ProviderEvent with BlockType/ToolUseId/ToolName/ToolInputJson, and add InputSchemaJson to ToolDefinition. All new fields are optional with null defaults so existing call sites compile without modification. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ts message history - Add ToolUse factory method to ProviderStreamEventFactory - Update AnthropicSdkStreamAdapter to detect ContentBlockStart/Stop and accumulate InputJsonDelta into tool_use events - Create AnthropicMessageBuilder to map ChatMessage[]/ProviderToolDefinition[] to Anthropic SDK params (MessageParam[], ToolUnion[]) - Update AnthropicProvider.StartStreamAsync to use message history, tools, and MaxTokens from ProviderRequest when Messages is set - Add ToolUseStreamAdapterTests covering factory, stream adapter, and builder behaviors Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…ge history - Update OpenAiMeaiStreamAdapter to detect FunctionCallContent in stream updates and emit ToolUse provider events - Add OpenAiMessageBuilder to map ProtocolChatMessage[]/ProviderToolDefinition[] to MEAI types (uses AIFunctionFactory.CreateDeclaration for tool schemas) - Update OpenAiCompatibleProvider to use message history and tools from ProviderRequest when present, falling back to single-prompt path - Add 5 new unit tests covering role mapping, FunctionCallContent, FunctionResultContent, tool building, and null-schema handling Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…olExecutor Implements ToolCallDispatcher that dispatches ProviderEvent tool-use requests through IToolExecutor, publishes ToolStartedEvent and ToolCompletedEvent runtime events, and returns a ContentBlock for feeding results back to the provider. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Add multi-iteration tool-calling loop that streams provider responses, detects tool-use events, dispatches them via ToolCallDispatcher, feeds results back as conversation messages, and resumes until the model stops requesting tools or MaxToolIterations is reached. - Create AgentLoopOptions for loop configuration (max iterations, max tokens) - Update ProviderBackedAgentKernel with tool-calling loop and backward-compat overload - Update AgentFrameworkBridge to accept IToolRegistry, build ToolExecutionContext, map tool definitions, and collect tool results - Add tool_call_roundtrip mock scenario to DeterministicMockModelProvider - Add parity test verifying the loop executes end-to-end Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

… context window management Introduces ConversationHistoryAssembler and ContextWindowManager to enable multi-turn conversations across session resume. Prior completed turns are read from the event store, assembled into ChatMessage[] pairs, truncated to a 100k-token budget, and threaded through PromptExecutionContext → AgentRunContext → ProviderBackedAgentKernel so the provider receives full context on every request. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Add NuGet package metadata including author, company, license, repository, and per-package descriptions to all publishable source projects. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

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

Adds SharpClawActivitySource and SharpClawMeterSource as central OTel instrumentation points; TurnActivityScope and ProviderActivityScope wrap turn/provider execution in Activity spans; NdjsonTraceFileSink writes completed spans as NDJSON for offline analysis. DefaultTurnRunner and ProviderBackedAgentKernel are wired to emit spans and record metrics. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Wraps IModelProvider registrations in ResilientProviderDecorator which retries transient failures (HttpRequestException 5xx, TaskCanceledException, IOException) with exponential backoff and jitter, honours Retry-After on 429 responses, and opens a circuit breaker after a configurable consecutive- failure threshold. All behaviour is driven by ProviderResilienceOptions and can be disabled via Enabled=false. Four unit tests cover retry-then-succeed, non-transient pass-through, circuit open, and probe-after-break-duration. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

This PR advances SharpClaw Code from a basic streaming wrapper to a more complete agent runtime by adding structured conversation history, a multi-iteration tool-calling loop across providers, resilience/observability primitives, plus initial CI/CD, docs, and examples.

Changes:

Introduces protocol/runtime support for multi-turn conversation history (ChatMessage/ContentBlock) and context-window truncation.
Implements provider tool-use extraction + a tool-calling loop (ProviderBackedAgentKernel + ToolCallDispatcher) with parity coverage.
Adds provider resilience (retry + circuit breaker) and baseline telemetry (OpenTelemetry spans + metrics) plus CI/release workflows and getting-started/integration docs & examples.

Reviewed changes

Copilot reviewed 78 out of 78 changed files in this pull request and generated 14 comments.

Show a summary per file

File	Description
tests/SharpClaw.Code.UnitTests/SharpClaw.Code.UnitTests.csproj	Adds Agents project reference so unit tests can cover new internal orchestration.
tests/SharpClaw.Code.UnitTests/Runtime/ConversationHistoryAssemblerTests.cs	Unit tests for assembling chat history from runtime events.
tests/SharpClaw.Code.UnitTests/Runtime/ContextWindowManagerTests.cs	Unit tests for token-budget truncation rules.
tests/SharpClaw.Code.UnitTests/Providers/ToolUseStreamAdapterTests.cs	Verifies streaming tool-use extraction and message/tool mapping for Anthropic + OpenAI-compatible adapters.
tests/SharpClaw.Code.UnitTests/Providers/ResilienceTests.cs	Tests retries and circuit breaker behavior for the resilience decorator.
tests/SharpClaw.Code.UnitTests/Protocol/ChatMessageSerializationTests.cs	Roundtrip serialization tests for new protocol message/content block types.
tests/SharpClaw.Code.UnitTests/Agents/ToolCallDispatcherTests.cs	Tests bridging provider tool-use events to tool execution + runtime events.
tests/SharpClaw.Code.ParityHarness/ParityScenarioTests.cs	Adds parity scenario to validate end-to-end tool-calling loop behavior.
tests/SharpClaw.Code.ParityHarness/ParityScenarioIds.cs	Registers new parity scenario ID.
tests/SharpClaw.Code.MockProvider/ParityProviderScenario.cs	Adds mock-provider scenario constant for tool roundtrip.
tests/SharpClaw.Code.MockProvider/DeterministicMockModelProvider.cs	Simulates tool-use then final response across loop iterations.
src/SharpClaw.Code.Web/SharpClaw.Code.Web.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Tools/SharpClaw.Code.Tools.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Tools/Models/ToolDefinition.cs	Extends tool definition to include optional input schema JSON.
src/SharpClaw.Code.Telemetry/SharpClaw.Code.Telemetry.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Telemetry/Metrics/SharpClawMeterSource.cs	Introduces central Meter + counters/histograms for runtime metrics.
src/SharpClaw.Code.Telemetry/Export/NdjsonTraceFileSink.cs	Adds NDJSON Activity span export sink for offline trace analysis.
src/SharpClaw.Code.Telemetry/Diagnostics/TurnActivityScope.cs	Adds Activity scope for turns with key tags and error reporting.
src/SharpClaw.Code.Telemetry/Diagnostics/SharpClawActivitySource.cs	Defines shared ActivitySource for SharpClaw tracing.
src/SharpClaw.Code.Telemetry/Diagnostics/ProviderActivityScope.cs	Adds Activity scope for provider calls with token tags.
src/SharpClaw.Code.Skills/SharpClaw.Code.Skills.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Sessions/SharpClaw.Code.Sessions.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Runtime/Turns/DefaultTurnRunner.cs	Wraps turn execution with tracing + metrics and propagates conversation history.
src/SharpClaw.Code.Runtime/SharpClaw.Code.Runtime.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Runtime/Context/PromptExecutionContext.cs	Adds conversation-history field to prompt execution context.
src/SharpClaw.Code.Runtime/Context/PromptContextAssembler.cs	Reads session event log and assembles/truncates conversation history for providers.
src/SharpClaw.Code.Runtime/Context/ConversationHistoryAssembler.cs	Implements event-log → ordered user/assistant message assembly.
src/SharpClaw.Code.Runtime/Context/ContextWindowManager.cs	Implements token-budget truncation for chat history.
src/SharpClaw.Code.Providers/SharpClaw.Code.Providers.csproj	Adds package description + internal visibility for unit tests.
src/SharpClaw.Code.Providers/Resilience/ResilientProviderDecorator.cs	Adds retry/backoff + basic circuit breaker around provider calls.
src/SharpClaw.Code.Providers/ProvidersServiceCollectionExtensions.cs	Wires resilience options + wraps providers conditionally.
src/SharpClaw.Code.Providers/OpenAiCompatibleProvider.cs	Adds support for request.Messages/tools/max tokens for MEAI-based provider.
src/SharpClaw.Code.Providers/Internal/ProviderStreamEventFactory.cs	Adds provider event factory method for tool-use events.
src/SharpClaw.Code.Providers/Internal/OpenAiMessageBuilder.cs	Maps SharpClaw chat/tool definitions to MEAI message/tool types.
src/SharpClaw.Code.Providers/Internal/OpenAiMeaiStreamAdapter.cs	Emits tool-use ProviderEvents from MEAI streaming function calls.
src/SharpClaw.Code.Providers/Internal/AnthropicSdkStreamAdapter.cs	Accumulates tool-use blocks from Anthropic streaming events.
src/SharpClaw.Code.Providers/Internal/AnthropicMessageBuilder.cs	Maps SharpClaw chat/tool definitions to Anthropic SDK message/tool types.
src/SharpClaw.Code.Providers/Configuration/ProviderResilienceOptions.cs	Adds resilience configuration options (retries, timeouts, circuit breaker).
src/SharpClaw.Code.Providers/AnthropicProvider.cs	Adds support for request.Messages/tools/max tokens via Anthropic SDK types.
src/SharpClaw.Code.Protocol/SharpClaw.Code.Protocol.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Protocol/Serialization/ProtocolJsonContext.cs	Adds source-gen JSON serialization support for new protocol models.
src/SharpClaw.Code.Protocol/Models/ProviderToolDefinition.cs	Introduces provider-facing tool definition contract.
src/SharpClaw.Code.Protocol/Models/ProviderRequest.cs	Extends provider request to include messages, tools, and max tokens.
src/SharpClaw.Code.Protocol/Models/ProviderEvent.cs	Extends provider events with tool-use metadata fields.
src/SharpClaw.Code.Protocol/Models/ContentBlock.cs	Introduces typed content blocks (text/tool_use/tool_result).
src/SharpClaw.Code.Protocol/Models/ChatMessage.cs	Introduces chat message model holding ordered content blocks.
src/SharpClaw.Code.Plugins/SharpClaw.Code.Plugins.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Permissions/SharpClaw.Code.Permissions.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Memory/SharpClaw.Code.Memory.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Mcp/SharpClaw.Code.Mcp.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Infrastructure/SharpClaw.Code.Infrastructure.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Git/SharpClaw.Code.Git.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Commands/SharpClaw.Code.Commands.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Cli/SharpClaw.Code.Cli.csproj	Adds NuGet package metadata description.
src/SharpClaw.Code.Agents/SharpClaw.Code.Agents.csproj	Adds options dependency, InternalsVisibleTo, and package description.
src/SharpClaw.Code.Agents/Services/AgentFrameworkBridge.cs	Maps tool registry → provider tools, builds tool execution context, and returns tool events/results.
src/SharpClaw.Code.Agents/Models/AgentRunContext.cs	Adds conversation history to agent run context.
src/SharpClaw.Code.Agents/Internal/ToolCallDispatcher.cs	Executes tool calls, publishes started/completed events, returns tool-result content blocks.
src/SharpClaw.Code.Agents/Internal/ProviderInvocationResult.cs	Extends provider invocation result to include tool results/events.
src/SharpClaw.Code.Agents/Internal/ProviderBackedAgentKernel.cs	Implements multi-iteration tool-calling loop and provider span/metrics.
src/SharpClaw.Code.Agents/Configuration/AgentLoopOptions.cs	Adds configuration for loop iterations and per-request token cap.
src/SharpClaw.Code.Agents/AgentsServiceCollectionExtensions.cs	Registers loop options and tool dispatcher in DI.
src/SharpClaw.Code.Acp/SharpClaw.Code.Acp.csproj	Adds NuGet package metadata description.
examples/WebApiAgent/WebApiAgent.csproj	Adds minimal ASP.NET Core agent example project.
examples/WebApiAgent/Program.cs	Implements minimal HTTP chat endpoint using SharpClaw runtime.
examples/WebApiAgent/appsettings.json	Provides example configuration for provider selection + logging.
examples/MinimalConsoleAgent/Program.cs	Adds minimal console agent example entry point.
examples/MinimalConsoleAgent/MinimalConsoleAgent.csproj	Adds minimal console agent example project file.
examples/MinimalConsoleAgent/appsettings.json	Provides example configuration for provider selection + logging.
examples/McpToolAgent/Program.cs	Adds example showing custom tool registration and invocation.
examples/McpToolAgent/McpToolAgent.csproj	Adds custom tool agent example project file.
examples/McpToolAgent/EchoTool.cs	Implements a simple custom echo tool using SharpClaw tool base class.
examples/McpToolAgent/appsettings.json	Provides example configuration for provider selection + logging.
docs/getting-started.md	Adds getting-started guide with build/run/config instructions.
docs/agent-framework-integration.md	Adds guide describing SharpClaw layering on Microsoft Agent Framework.
Directory.Build.props	Adds shared NuGet metadata (authors/license/repo URLs).
.github/workflows/release.yml	Adds tag-driven NuGet package publish workflow.
.github/workflows/ci.yml	Adds multi-OS build/test workflow with coverage artifact upload.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-11T19:11:00Z

src/SharpClaw.Code.Agents/Internal/ProviderBackedAgentKernel.cs

+            var messages = new List<ChatMessage>();
+            if (!string.IsNullOrWhiteSpace(request.Instructions))
+            {
+                messages.Add(new ChatMessage("system", [new ContentBlock(ContentBlockKind.Text, request.Instructions, null, null, null, null)]));
+            }


The tool-calling loop prepends a ChatMessage with role == "system" and also passes the same instructions via ProviderRequest.SystemPrompt. Anthropic only supports user/assistant roles and AnthropicMessageBuilder maps any non-assistant role to user, so this will turn system instructions into an extra user message (and potentially duplicate the system prompt). Consider removing the explicit "system" ChatMessage from the shared message list and letting each provider adapter apply SystemPrompt appropriately (or explicitly filtering out system-role messages for Anthropic).

Suggested change

var messages = new List<ChatMessage>();

if (!string.IsNullOrWhiteSpace(request.Instructions))

{

messages.Add(new ChatMessage("system", [new ContentBlock(ContentBlockKind.Text, request.Instructions, null, null, null, null)]));

}

// Do not add request.Instructions as a shared "system" chat message here.

// Provider adapters should apply system instructions via ProviderRequest.SystemPrompt

// so providers that do not support a native "system" role (for example Anthropic)

// do not receive duplicated or remapped instruction turns.

var messages = new List<ChatMessage>();

Copilot · 2026-04-11T19:11:01Z

src/SharpClaw.Code.Agents/Internal/ProviderBackedAgentKernel.cs


-            await foreach (var providerEvent in stream.Events.WithCancellation(cancellationToken))
+            for (var iteration = 0; iteration < options.MaxToolIterations; iteration++)
            {


terminalUsage is declared outside the iteration loop and never reset per iteration. If a later provider call returns a terminal event without Usage (or Usage=null), the loop will reuse the previous iteration's usage when setting ProviderActivityScope tags and when returning the final UsageSnapshot, producing incorrect telemetry/usage reporting. Track usage per iteration (reset to null at loop start) and only promote it to the overall terminal usage when the current iteration provides it.

Suggested change

{

{

terminalUsage = null;

Copilot · 2026-04-11T19:11:01Z

src/SharpClaw.Code.Agents/Internal/ProviderBackedAgentKernel.cs

+                    providerSw.Stop();
+                    providerScope.SetCompleted(terminalUsage?.InputTokens, terminalUsage?.OutputTokens);
+                    SharpClawMeterSource.ProviderDuration.Record(providerSw.Elapsed.TotalMilliseconds);


ProviderActivityScope.SetCompleted is called with terminalUsage which can reflect a previous iteration (since terminalUsage isn't iteration-scoped). This can attach incorrect token counts to spans for calls that didn't emit usage. Use an iteration-local usage variable for span tagging, and separately maintain the final/last-known usage for the overall ProviderInvocationResult.

Copilot · 2026-04-11T19:11:01Z

src/SharpClaw.Code.Runtime/Context/ContextWindowManager.cs

+    /// <list type="number">
+    ///   <item>The system message (role == "system") is always kept.</item>
+    ///   <item>The most-recent user message is always kept.</item>
+    ///   <item>Oldest non-system messages are dropped until the budget is satisfied.</item>
+    /// </list>


The XML docs say the "most-recent user message" is always kept, but the implementation actually guarantees only the most recent non-system message is kept (which may be an assistant message, and the code does not look for the most recent user role specifically). Either update the documentation to match the behavior or adjust the truncation logic to explicitly preserve the most recent user message.

Copilot · 2026-04-11T19:11:01Z

src/SharpClaw.Code.Agents/Internal/ToolCallDispatcher.cs

+        var toolName = toolUseEvent.ToolName ?? string.Empty;
+        var toolInputJson = toolUseEvent.ToolInputJson ?? "{}";
+        var toolUseId = toolUseEvent.ToolUseId ?? string.Empty;


Defaulting ToolName/ToolUseId to empty strings can lead to publishing ToolStartedEvent/ToolCompletedEvent with empty identifiers and then invoking the tool executor with an empty tool name. It would be safer to validate these fields on toolUseEvent and return an error tool-result block (and/or a classified runtime event) without calling the executor when required metadata is missing.

Suggested change

var toolName = toolUseEvent.ToolName ?? string.Empty;

var toolInputJson = toolUseEvent.ToolInputJson ?? "{}";

var toolUseId = toolUseEvent.ToolUseId ?? string.Empty;

global::System.ArgumentException.ThrowIfNullOrWhiteSpace(toolUseEvent.ToolName);

global::System.ArgumentException.ThrowIfNullOrWhiteSpace(toolUseEvent.ToolUseId);

var toolName = toolUseEvent.ToolName;

var toolInputJson = toolUseEvent.ToolInputJson ?? "{}";

var toolUseId = toolUseEvent.ToolUseId;

Copilot · 2026-04-11T19:11:03Z

docs/agent-framework-integration.md

+   - Placeholder response when stream is empty
+
+**Loop Configuration:** Controlled by `AgentLoopOptions`:
+- `MaxToolIterations` — maximum rounds (default ~10)


The documented default for AgentLoopOptions.MaxToolIterations is "~10", but the implementation sets the default to 25. Please update the doc to reflect the actual default (or update AgentLoopOptions if 10 is the intended default).

Suggested change

- `MaxToolIterations` — maximum rounds (default ~10)

- `MaxToolIterations` — maximum rounds (default 25)

Copilot · 2026-04-11T19:11:03Z

docs/agent-framework-integration.md

+public interface ISharpClawTool
+{
+    string Name { get; }
+    string Description { get; }
+    string InputSchemaJson { get; }
+    Task<string> ExecuteAsync(string input, ToolExecutionContext context, CancellationToken cancellationToken);
+}


The ISharpClawTool interface shown here does not match the actual ISharpClawTool contract in src/SharpClaw.Code.Tools/Abstractions/ISharpClawTool.cs (which exposes ToolDefinition and ExecuteAsync(context, request)). This will mislead tool implementers. Please update the snippet to mirror the real interface and tool implementation pattern used in the codebase (e.g., SharpClawToolBase + ToolDefinition).

Copilot · 2026-04-11T19:11:03Z

examples/MinimalConsoleAgent/appsettings.json

+    "Provider": "Anthropic",
+    "Anthropic": {
+      "ApiKey": "<YOUR_ANTHROPIC_API_KEY>",
+      "Model": "claude-sonnet-4-5"


This example appsettings.json uses "SharpClaw:Provider" and "SharpClaw:Anthropic" keys, but the runtime binds provider options under "SharpClaw:Providers:*" (see ProvidersServiceCollectionExtensions). As-is, these settings won’t be picked up and the example likely won’t authenticate/configure the provider. Update to the correct section names (e.g., SharpClaw:Providers:Anthropic plus SharpClaw:Providers:Catalog:DefaultProvider).

Suggested change

"Provider": "Anthropic",

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

"Providers": {

"Catalog": {

"DefaultProvider": "Anthropic"

},

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

}

Copilot · 2026-04-11T19:11:03Z

examples/WebApiAgent/appsettings.json

+    "Provider": "Anthropic",
+    "Anthropic": {
+      "ApiKey": "<YOUR_ANTHROPIC_API_KEY>",
+      "Model": "claude-sonnet-4-5"


This example appsettings.json uses "SharpClaw:Provider" and "SharpClaw:Anthropic" keys, but the runtime binds provider options under "SharpClaw:Providers:*" (see ProvidersServiceCollectionExtensions). As-is, these settings won’t be picked up and the example likely won’t authenticate/configure the provider. Update to the correct section names (e.g., SharpClaw:Providers:Anthropic plus SharpClaw:Providers:Catalog:DefaultProvider).

Suggested change

"Provider": "Anthropic",

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

"Providers": {

"Catalog": {

"DefaultProvider": "Anthropic"

},

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

}

Copilot · 2026-04-11T19:11:04Z

examples/McpToolAgent/appsettings.json

+    "Provider": "Anthropic",
+    "Anthropic": {
+      "ApiKey": "<YOUR_ANTHROPIC_API_KEY>",
+      "Model": "claude-sonnet-4-5"


This example appsettings.json uses "SharpClaw:Provider" and "SharpClaw:Anthropic" keys, but the runtime binds provider options under "SharpClaw:Providers:*" (see ProvidersServiceCollectionExtensions). As-is, these settings won’t be picked up and the example likely won’t authenticate/configure the provider. Update to the correct section names (e.g., SharpClaw:Providers:Anthropic plus SharpClaw:Providers:Catalog:DefaultProvider).

Suggested change

"Provider": "Anthropic",

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

"Providers": {

"Catalog": {

"DefaultProvider": "Anthropic"

},

"Anthropic": {

"ApiKey": "<YOUR_ANTHROPIC_API_KEY>",

"Model": "claude-sonnet-4-5"

}

Telli and others added 14 commits April 10, 2026 21:48

build: add NuGet package metadata for publishing

2d7a1f1

Add NuGet package metadata including author, company, license, repository, and per-package descriptions to all publishable source projects. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

ci: add GitHub Actions workflows for CI and NuGet release

f89906c

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

docs: add getting started guide

8c27ca4

docs: add Microsoft Agent Framework integration guide

0fbddbf

docs: add example projects (minimal console, web API, custom tool agent)

2a02159

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

Copilot AI review requested due to automatic review settings April 11, 2026 19:02

Copilot started reviewing on behalf of Telli April 11, 2026 19:03 View session

Copilot AI reviewed Apr 11, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Implement Phase 1: tool-calling loop, conversation history, CI/CD, docs, resilience, observability#3

Implement Phase 1: tool-calling loop, conversation history, CI/CD, docs, resilience, observability#3
Telli wants to merge 14 commits intomainfrom
phase1-implementation

Telli commented Apr 11, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Copilot AI Apr 11, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	- `MaxToolIterations` — maximum rounds (default ~10)
	- `MaxToolIterations` — maximum rounds (default 25)

Conversation

Telli commented Apr 11, 2026

Summary

P0 — Tool-Calling Loop + Conversation History

P1 — Infrastructure

P2 — Production Hardening

Stats

Test plan

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 11, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants