fix: restore matrix example and update actions

.github/workflows/ci.yml

@@ -15,7 +15,7 @@ jobs:
     name: Lint
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2
@@ -40,7 +40,7 @@ jobs:
     name: Type Check
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2
@@ -62,7 +62,7 @@ jobs:
     name: Unit Tests
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2
@@ -84,7 +84,7 @@ jobs:
     name: Smoke Test
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2

.github/workflows/codeql.yml

@@ -25,18 +25,18 @@ jobs:
 
     steps:
       - name: Checkout repository
-        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4
+        uses: github/codeql-action/init@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
         with:
           languages: ${{ matrix.language }}
           queries: security-extended,security-and-quality
 
       - name: Autobuild
-        uses: github/codeql-action/autobuild@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4
+        uses: github/codeql-action/autobuild@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4
+        uses: github/codeql-action/analyze@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
         with:
           category: "/language:${{ matrix.language }}"

.github/workflows/dependency-review.yml

@@ -14,7 +14,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
-        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Dependency Review
         uses: actions/dependency-review-action@a1d282b36b6f3519aa1f3fc636f609c47dddb294 # v4

.github/workflows/docs.yml

@@ -27,7 +27,7 @@ jobs:
   build:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2

.github/workflows/hero-tests.yml

@@ -29,7 +29,7 @@ jobs:
 
     steps:
       - name: Checkout
-        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
         with:
           ref: ${{ github.head_ref || github.ref }}
           token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/integration.yml

@@ -22,7 +22,7 @@ jobs:
       id-token: write
       contents: read
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2

.github/workflows/release.yml

@@ -30,7 +30,7 @@ jobs:
       tag_exists: ${{ steps.calculate_version.outputs.tag_exists }}
       pypi_exists: ${{ steps.calculate_version.outputs.pypi_exists }}
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
         with:
           fetch-depth: 0  # Fetch all history and tags
 
@@ -179,7 +179,7 @@ jobs:
     runs-on: ubuntu-latest
     needs: version
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Install uv
         uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2
@@ -240,7 +240,7 @@ jobs:
     permissions:
       contents: write
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
         with:
           fetch-depth: 0
 
@@ -275,7 +275,7 @@ jobs:
           path: dist/
 
       - name: Publish to PyPI
-        uses: pypa/gh-action-pypi-publish@cef221092ed1bacb1cc03d23a2d87d1d172e277b # release/v1
+        uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2
 
   github-release:
     name: Create GitHub Release
@@ -285,7 +285,7 @@ jobs:
     permissions:
       contents: write
     steps:
-      - uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7
 
       - name: Download build artifacts
         uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8

examples/README.md

@@ -1,75 +1,68 @@
 # Examples
 
-Working examples are in `tests/integration/` — they are the best reference for how to use pytest-skill-engineering.
+Working examples live in [`tests/integration/copilot/`](../tests/integration/copilot/). They use real GitHub Copilot models and are the best reference for pytest-skill-engineering.
 
 ## Test Files
 
 | File | What it demonstrates |
-|------|---------------------|
-| [pydantic/test_01_basic.py](../tests/integration/pydantic/test_01_basic.py) | Natural language → tool usage with Banking and Todo servers |
-| [pydantic/test_02_models.py](../tests/integration/pydantic/test_02_models.py) | Model comparison (parametrize) |
-| [pydantic/test_03_prompts.py](../tests/integration/pydantic/test_03_prompts.py) | System prompt comparison |
-| [pydantic/test_04_matrix.py](../tests/integration/pydantic/test_04_matrix.py) | Model × prompt 2×2 grid |
-| [pydantic/test_05_skills.py](../tests/integration/pydantic/test_05_skills.py) | Skills with references and metadata |
-| [pydantic/test_06_sessions.py](../tests/integration/pydantic/test_06_sessions.py) | Multi-turn session continuity |
-| [pydantic/test_07_clarification.py](../tests/integration/pydantic/test_07_clarification.py) | ClarificationDetection feature |
-| [pydantic/test_08_scoring.py](../tests/integration/pydantic/test_08_scoring.py) | llm_score + ScoringDimension |
-| [pydantic/test_09_cli.py](../tests/integration/pydantic/test_09_cli.py) | CLIServer wrapping shell commands |
-| [pydantic/test_10_ab_servers.py](../tests/integration/pydantic/test_10_ab_servers.py) | A/B server comparison |
-| [pydantic/test_11_iterations.py](../tests/integration/pydantic/test_11_iterations.py) | --aitest-iterations=N reliability |
-| [pydantic/test_12_custom_agents.py](../tests/integration/pydantic/test_12_custom_agents.py) | Eval.from_agent_file + load_custom_agent |
+|------|----------------------|
+| [`test_01_basic.py`](../tests/integration/copilot/test_01_basic.py) | Basic file creation and refactoring |
+| [`test_02_models.py`](../tests/integration/copilot/test_02_models.py) | Model comparison |
+| [`test_03_instructions.py`](../tests/integration/copilot/test_03_instructions.py) | System prompt comparison and tool restrictions |
+| [`test_04_matrix.py`](../tests/integration/copilot/test_04_matrix.py) | Model × system prompt matrix |
+| [`test_05_skills.py`](../tests/integration/copilot/test_05_skills.py) | Skill A/B comparison |
+| [`test_06_sessions.py`](../tests/integration/copilot/test_06_sessions.py) | Multi-turn sessions |
+| [`test_07_clarification.py`](../tests/integration/copilot/test_07_clarification.py) | Clarification detection |
+| [`test_08_scoring.py`](../tests/integration/copilot/test_08_scoring.py) | LLM scoring |
+| [`test_09_cli.py`](../tests/integration/copilot/test_09_cli.py) | CLI workflows |
+| [`test_10_ab_servers.py`](../tests/integration/copilot/test_10_ab_servers.py) | Configuration A/B comparison |
+| [`test_11_iterations.py`](../tests/integration/copilot/test_11_iterations.py) | Iteration reliability |
+| [`test_12_custom_agents.py`](../tests/integration/copilot/test_12_custom_agents.py) | Custom agent dispatch |
+| [`test_13_plugins.py`](../tests/integration/copilot/test_13_plugins.py) | Plugin discovery and loading |
+| [`test_14_skill_evals.py`](../tests/integration/copilot/test_14_skill_evals.py) | Skill eval execution |
+| [`test_15_skill_refinement.py`](../tests/integration/copilot/test_15_skill_refinement.py) | Skill refinement |
+| [`test_16_skill_benchmark.py`](../tests/integration/copilot/test_16_skill_benchmark.py) | Skill benchmarking |
+| [`test_17_plugin_skill_workflow.py`](../tests/integration/copilot/test_17_plugin_skill_workflow.py) | End-to-end plugin skill workflow |
+| [`test_events.py`](../tests/integration/copilot/test_events.py) | Copilot SDK event capture |
 
 ## Run Examples
 
 ```bash
-# Prerequisites
-uv add pytest-skill-engineering
-az login               # For Azure OpenAI
+# Authenticate once
+gh auth login
 
 # Run basic usage tests
-uv run python -m pytest tests/integration/pydantic/test_01_basic.py -v
+uv run python -m pytest tests/integration/copilot/test_01_basic.py -v
 
-# Run all pydantic tests
-uv run python -m pytest tests/integration/pydantic/ -v
+# Run all Copilot integration tests
+uv run python -m pytest tests/integration/copilot/ -v
 
-# Run with report
-uv run python -m pytest tests/integration/pydantic/ -v --aitest-html=report.html
+# Generate a report
+uv run python -m pytest tests/integration/copilot/test_04_matrix.py -v \
+  --aitest-html=report.html
 ```
 
-## Test Servers
+## MCP Server Example
 
-Two built-in test servers for natural language testing:
+`CopilotEval` accepts MCP server configurations in the Copilot SDK format:
 
-### Banking Server
 ```python
-@pytest.fixture(scope="module")
-def banking_server():
-    return MCPServer(
-        command=["python", "-m", "pytest_skill_engineering.testing.banking_mcp"],
-        wait=Wait.for_tools(["get_balance", "transfer"]),
-    )
-
-
-# Test: "What's my checking balance?"
+import sys
+
+from pytest_skill_engineering.copilot import CopilotEval
+
+
+banking_eval = CopilotEval(
+    name="banking",
+    instructions="Use the banking tools for every account request.",
+    mcp_servers={
+        "banking": {
+            "command": sys.executable,
+            "args": ["-m", "pytest_skill_engineering.testing.banking_mcp"],
+            "tools": ["*"],
+        }
+    },
+)
 ```
 
-### Todo Server
-```python
-@pytest.fixture(scope="module")
-def todo_server():
-    return MCPServer(
-        command=["python", "-m", "pytest_skill_engineering.testing.todo_mcp"],
-        wait=Wait.for_tools(["add_task", "list_tasks"]),
-    )
-
-
-# Test: "Add buy milk to my shopping list"
-```
-
-## Fixtures
-
-See [conftest.py](../tests/integration/conftest.py) for fixture patterns:
-
-- `banking_server` / `todo_server` — Test MCP servers
-- `DEFAULT_MODEL`, `DEFAULT_RPM`, `DEFAULT_TPM`, `DEFAULT_MAX_TURNS` — Constants for eval creation
-- Azure token handling
+The package includes banking and todo MCP servers for integration scenarios. See [`tests/showcase/`](../tests/showcase/) for a complete banking example.

tests/integration/README.md

@@ -16,6 +16,7 @@ tests/integration/
     ├── test_01_basic.py         # Basic file creation and refactoring
     ├── test_02_models.py        # Model comparison
     ├── test_03_instructions.py  # System prompt and tool filtering
+    ├── test_04_matrix.py        # Model × system prompt matrix
     ├── test_05_skills.py        # Skill A/B comparison
     ├── test_06_sessions.py      # Multi-turn sessions
     ├── test_07_clarification.py # Clarification detection

tests/integration/copilot/test_04_matrix.py

@@ -0,0 +1,64 @@
+"""Level 04 — Model × system prompt matrix comparison.
+
+Runs the same task across every configured model and two distinct system
+prompts. The report exposes both dimensions for side-by-side comparison.
+
+Run with: pytest tests/integration/copilot/test_04_matrix.py -v
+"""
+
+from __future__ import annotations
+
+import pytest
+
+from pytest_skill_engineering.copilot.eval import CopilotEval
+
+from .conftest import MODELS
+
+pytestmark = [pytest.mark.copilot]
+
+SYSTEM_PROMPTS = {
+    "typed": (
+        "Write fully typed Python. Every function must have parameter and return "
+        "type annotations plus a concise docstring."
+    ),
+    "minimal": (
+        "Write minimal Python. Do not add type annotations, docstrings, comments, "
+        "tests, or supporting files."
+    ),
+}
+
+
+class TestModelSystemPromptMatrix:
+    """Compare each model under typed and minimal system prompts."""
+
+    @pytest.mark.parametrize("model", MODELS)
+    @pytest.mark.parametrize("system_prompt_name", SYSTEM_PROMPTS)
+    async def test_greeting_module(
+        self,
+        copilot_eval,
+        tmp_path,
+        model: str,
+        system_prompt_name: str,
+    ) -> None:
+        """Each matrix cell creates the requested module in its required style."""
+        agent = CopilotEval(
+            name=f"{model}-{system_prompt_name}",
+            model=model,
+            instructions=SYSTEM_PROMPTS[system_prompt_name],
+            working_directory=str(tmp_path),
+        )
+
+        result = await copilot_eval(
+            agent,
+            "Create greeting.py with greet(name), returning 'Hello, {name}!'.",
+        )
+
+        assert result.success, result.error
+        content = (tmp_path / "greeting.py").read_text(encoding="utf-8")
+        assert "def greet" in content
+        if system_prompt_name == "typed":
+            assert "-> str" in content
+            assert '"""' in content or "'''" in content
+        else:
+            assert "->" not in content
+            assert '"""' not in content and "'''" not in content