Add full e2e test: define -> implement -> execute workflow

Update the claude-code-e2e job to test the COMPLETE DeepWork workflow:
1. /deepwork_jobs.define - Create a new job from scratch
2. /deepwork_jobs.implement - Generate step instruction files
3. /fruits.identify - Execute the generated identify command
4. /fruits.classify - Execute the generated classify command

This tests the actual user experience rather than just validating
pre-existing fixtures. The test provides deterministic instructions
for creating a 'fruits' job with identify and classify steps.

Author: nhorton

.github/workflows/claude-code-test.yml

@@ -26,7 +26,7 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  # Job 1: Validate command generation (always runs)
+  # Job 1: Validate command generation from fixtures (always runs, no API key needed)
   validate-generation:
     runs-on: ubuntu-latest
     steps:
@@ -86,7 +86,8 @@ jobs:
 
           echo "Command generation validated successfully!"
 
-  # Job 2: End-to-end test with Claude Code (only when API key is available)
+  # Job 2: Full end-to-end test with Claude Code
+  # Tests the COMPLETE workflow: define job -> implement -> execute
   claude-code-e2e:
     runs-on: ubuntu-latest
     needs: validate-generation
@@ -132,91 +133,221 @@ jobs:
         if: steps.check-key.outputs.has_key == 'true'
         run: uv sync
 
-      - name: Set up test project
+      - name: Set up fresh test project
         if: steps.check-key.outputs.has_key == 'true'
         run: |
-          mkdir -p test_project/.deepwork/jobs
-          mkdir -p test_project/.claude  # Required for platform detection
-          cp -r tests/fixtures/jobs/fruits test_project/.deepwork/jobs/
+          # Create a fresh project with NO pre-existing job definitions
+          mkdir -p test_project/.claude
 
           cd test_project
           git init
           git config user.email "test@test.com"
           git config user.name "Test"
-          echo "# CI Test Project" > README.md
+          echo "# CI Test Project - DeepWork E2E Test" > README.md
           git add . && git commit -m "init"
           cd ..
 
-          # Run deepwork install to set up the project (this also runs sync)
+          # Install deepwork (this sets up .deepwork/ with standard jobs only)
           uv run deepwork install --platform claude --path test_project
 
-          echo "Test project setup complete"
+          echo "Fresh test project setup complete"
+          echo "Available commands:"
           ls -la test_project/.claude/commands/
 
-      - name: Run Claude Code - Identify Step
+      # STEP 1: Use /deepwork_jobs.define to CREATE the fruits job
+      - name: Create job with /deepwork_jobs.define
+        if: steps.check-key.outputs.has_key == 'true'
+        working-directory: test_project
+        timeout-minutes: 10
+        run: |
+          echo "=== Running /deepwork_jobs.define to create fruits job ==="
+
+          # Provide detailed, deterministic instructions for creating the job
+          claude --yes --print "/deepwork_jobs.define" <<'PROMPT_EOF'
+          I want to create a simple job called "fruits" for identifying and classifying fruits.
+
+          Here are the EXACT specifications - please create the job.yml with these exact details:
+
+          Job name: fruits
+          Version: 1.0.0
+          Summary: Identify and classify fruits from a mixed list of items
+
+          Description: A simple workflow that takes a list of mixed items, identifies which are fruits, then classifies them by category. Designed for CI testing.
+
+          Steps:
+          1. Step ID: identify
+             Name: Identify Fruits
+             Description: Filter a list of items to identify only the fruits
+             Input: raw_items (user parameter) - A comma-separated list of items
+             Output: identified_fruits.md
+             Dependencies: none
+
+          2. Step ID: classify
+             Name: Classify Fruits
+             Description: Organize identified fruits into categories (citrus, tropical, berries, etc.)
+             Input: identified_fruits.md (file from step identify)
+             Output: classified_fruits.md
+             Dependencies: identify
+
+          Please create this job definition now. Do not ask questions - use these exact specifications.
+          PROMPT_EOF
+
+          # Verify the job.yml was created
+          echo "=== Checking job.yml was created ==="
+          if [ -f ".deepwork/jobs/fruits/job.yml" ]; then
+            echo "SUCCESS: job.yml created"
+            cat .deepwork/jobs/fruits/job.yml
+          else
+            echo "ERROR: job.yml was not created"
+            echo "Contents of .deepwork/jobs/:"
+            ls -la .deepwork/jobs/ || echo "No jobs directory"
+            exit 1
+          fi
+
+      # STEP 2: Use /deepwork_jobs.implement to generate step instructions
+      - name: Generate step instructions with /deepwork_jobs.implement
+        if: steps.check-key.outputs.has_key == 'true'
+        working-directory: test_project
+        timeout-minutes: 10
+        run: |
+          echo "=== Running /deepwork_jobs.implement to generate step instructions ==="
+
+          claude --yes --print "/deepwork_jobs.implement" <<'PROMPT_EOF'
+          Please implement the "fruits" job that was just defined.
+
+          For the identify step, create instructions that:
+          - Parse the comma-separated raw_items input
+          - Identify which items are fruits (apple, banana, orange, mango, grape, etc.)
+          - Output a markdown file listing the identified fruits
+
+          For the classify step, create instructions that:
+          - Read identified_fruits.md from the previous step
+          - Classify fruits into categories: Citrus (orange, lemon), Tropical (banana, mango), Pome (apple, pear), Berries, etc.
+          - Output a markdown file with fruits organized by category
+
+          Generate the step instruction files now.
+          PROMPT_EOF
+
+          # Verify step files were created
+          echo "=== Checking step files were created ==="
+          if [ -f ".deepwork/jobs/fruits/steps/identify.md" ] && [ -f ".deepwork/jobs/fruits/steps/classify.md" ]; then
+            echo "SUCCESS: Step instruction files created"
+            echo "--- identify.md ---"
+            cat .deepwork/jobs/fruits/steps/identify.md
+            echo ""
+            echo "--- classify.md ---"
+            cat .deepwork/jobs/fruits/steps/classify.md
+          else
+            echo "ERROR: Step files were not created"
+            ls -la .deepwork/jobs/fruits/steps/ || echo "No steps directory"
+            exit 1
+          fi
+
+          # Run sync to generate the slash commands
+          echo "=== Running deepwork sync to generate commands ==="
+          cd ..
+          uv run deepwork sync --path test_project
+
+          echo "=== Checking generated commands ==="
+          ls -la test_project/.claude/commands/
+
+          if [ -f "test_project/.claude/commands/fruits.identify.md" ] && [ -f "test_project/.claude/commands/fruits.classify.md" ]; then
+            echo "SUCCESS: Slash commands generated"
+          else
+            echo "ERROR: Slash commands were not generated"
+            exit 1
+          fi
+
+      # STEP 3: Execute the generated /fruits.identify command
+      - name: Run /fruits.identify
         if: steps.check-key.outputs.has_key == 'true'
         working-directory: test_project
         timeout-minutes: 5
         run: |
-          # Run the identify step with a deterministic input
-          # Using --print to output result, --yes to auto-accept
-          claude --yes --print "/fruits.identify" <<EOF
-          raw_items: apple, car, banana, chair, orange, table, mango, laptop
-          EOF
+          echo "=== Running /fruits.identify with test input ==="
+
+          claude --yes --print "/fruits.identify" <<'PROMPT_EOF'
+          raw_items: apple, car, banana, chair, orange, table, mango, laptop, grape, bicycle
+          PROMPT_EOF
 
           # Verify output was created
           if [ -f "identified_fruits.md" ]; then
-            echo "Identify step completed successfully!"
+            echo "SUCCESS: identified_fruits.md created"
             echo "--- Output ---"
             cat identified_fruits.md
           else
             echo "ERROR: identified_fruits.md was not created"
             exit 1
           fi
 
-      - name: Run Claude Code - Classify Step
+      # STEP 4: Execute the generated /fruits.classify command
+      - name: Run /fruits.classify
         if: steps.check-key.outputs.has_key == 'true'
         working-directory: test_project
         timeout-minutes: 5
         run: |
-          # Run the classify step
+          echo "=== Running /fruits.classify ==="
+
           claude --yes --print "/fruits.classify"
 
           # Verify output was created
           if [ -f "classified_fruits.md" ]; then
-            echo "Classify step completed successfully!"
+            echo "SUCCESS: classified_fruits.md created"
             echo "--- Output ---"
             cat classified_fruits.md
           else
             echo "ERROR: classified_fruits.md was not created"
             exit 1
           fi
 
-      - name: Validate outputs
+      # STEP 5: Validate the complete workflow output
+      - name: Validate complete workflow
         if: steps.check-key.outputs.has_key == 'true'
         working-directory: test_project
         run: |
-          echo "=== Validating outputs ==="
+          echo "=== Validating complete workflow ==="
 
           # Check identified_fruits.md contains expected fruits
           echo "Checking identified_fruits.md..."
           grep -qi "apple" identified_fruits.md || (echo "Missing: apple" && exit 1)
           grep -qi "banana" identified_fruits.md || (echo "Missing: banana" && exit 1)
           grep -qi "orange" identified_fruits.md || (echo "Missing: orange" && exit 1)
           grep -qi "mango" identified_fruits.md || (echo "Missing: mango" && exit 1)
+          grep -qi "grape" identified_fruits.md || (echo "Missing: grape" && exit 1)
+          echo "  ✓ All expected fruits found in identified_fruits.md"
 
           # Check classified_fruits.md has expected structure
           echo "Checking classified_fruits.md..."
-          grep -qi "citrus\|tropical\|pome" classified_fruits.md || (echo "Missing fruit categories" && exit 1)
-
-          echo "All validations passed!"
+          grep -qi "citrus\|tropical\|pome\|berr" classified_fruits.md || (echo "Missing fruit categories" && exit 1)
+          echo "  ✓ Fruit categories found in classified_fruits.md"
+
+          # Verify job structure was created correctly
+          echo "Checking job structure..."
+          test -f .deepwork/jobs/fruits/job.yml || (echo "Missing job.yml" && exit 1)
+          test -f .deepwork/jobs/fruits/steps/identify.md || (echo "Missing identify.md" && exit 1)
+          test -f .deepwork/jobs/fruits/steps/classify.md || (echo "Missing classify.md" && exit 1)
+          echo "  ✓ Job structure is complete"
+
+          echo ""
+          echo "=========================================="
+          echo "  ALL E2E TESTS PASSED SUCCESSFULLY!"
+          echo "=========================================="
+          echo ""
+          echo "Workflow tested:"
+          echo "  1. /deepwork_jobs.define - Created job specification"
+          echo "  2. /deepwork_jobs.implement - Generated step instructions"
+          echo "  3. /fruits.identify - Executed identify step"
+          echo "  4. /fruits.classify - Executed classify step"
+          echo ""
 
       - name: Upload test artifacts
         if: steps.check-key.outputs.has_key == 'true' && always()
         uses: actions/upload-artifact@v4
         with:
-          name: claude-code-test-outputs
+          name: claude-code-e2e-outputs
           path: |
+            test_project/.deepwork/jobs/fruits/
+            test_project/.claude/commands/fruits.*.md
             test_project/identified_fruits.md
             test_project/classified_fruits.md
           retention-days: 7