Test Plugin

Comprehensive test of all Bluera Base plugin functionality (hooks + slash commands).

⚠️ EXCLUDED COMMANDS

The following commands are NOT tested by this suite:

/clean

~/.claude

DO NOT ADD

/clean

Context

!

echo "=== Bluera Base Plugin Test ===" && echo "Hooks: $(ls "${CLAUDE_PLUGIN_ROOT:-$(pwd)}/hooks/"*.sh 2>/dev/null | wc -l | tr -d ' ') files" && echo "Skills: $(ls -d "${CLAUDE_PLUGIN_ROOT:-$(pwd)}/skills/"*/ 2>/dev/null | wc -l | tr -d ' ') directories" && echo "Commands: $(ls "${CLAUDE_PLUGIN_ROOT:-$(pwd)}/commands/"*.md 2>/dev/null | wc -l | tr -d ' ') files"

Pre-Test Cleanup

Remove any leftover artifacts from previous test runs:

rm -rf .bluera/bluera-base/state/milhouse-loop.md
rm -rf /tmp/bluera-base-test

Test Content Setup

Create a temporary test directory with various project types:

mkdir -p /tmp/bluera-base-test
cd /tmp/bluera-base-test
git init
echo '{"name": "test-project", "version": "1.0.0"}' > package.json
echo 'console.log("test");' > index.js
git add .
git commit -m "initial"

Workflow

Execute each test in order. Mark each as PASS or FAIL.

Part 1: Hook Registration

1. Hook File Structure: Verify hooks.json has expected structure

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   cat "$PLUGIN_PATH/hooks/hooks.json" | jq -e '.hooks.PreToolUse and .hooks.PostToolUse and .hooks.Stop and .hooks.Notification'
   

   • Expected: Returns

     true

     (all hook types registered)
   • PASS if command succeeds with truthy output

2. Hook Scripts Exist: Verify all referenced hooks have scripts

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   MISSING=0
   for script in $(grep -oE '[a-z-]+\.sh' "$PLUGIN_PATH/hooks/hooks.json" | sort -u); do
     if [[ ! -f "$PLUGIN_PATH/hooks/$script" ]]; then
       echo "MISSING: $script"
       MISSING=$((MISSING + 1))
     fi
   done
   TOTAL=$(grep -oE '[a-z-]+\.sh' "$PLUGIN_PATH/hooks/hooks.json" | sort -u | wc -l | tr -d ' ')
   echo "Checked $TOTAL referenced scripts, $MISSING missing"
   [[ $MISSING -eq 0 ]] && echo "All hook scripts present"
   

   • Expected: All scripts referenced in hooks.json exist
   • PASS if output shows "All hook scripts present"

Part 2: PreToolUse Hook (block-manual-release.sh)

1. Block npm version: Test that manual npm version is blocked

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"tool_input": {"command": "npm version patch"}}' | bash "$PLUGIN_PATH/hooks/block-manual-release.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Message about using /bluera-base:release, exit code 2
   • PASS if output contains "Use /bluera-base:release" and exits 2

2. Block git tag: Test that manual git tagging is blocked

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"tool_input": {"command": "git tag v1.0.0"}}' | bash "$PLUGIN_PATH/hooks/block-manual-release.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Blocked with exit code 2
   • PASS if exits 2

3. Allow skill-prefixed release: Test that /bluera-base:release skill can run version commands

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"tool_input": {"command": "__SKILL__=release npm version patch"}}' | bash "$PLUGIN_PATH/hooks/block-manual-release.sh"
   echo "Exit code: $?"
   

   • Expected: Allowed (exit code 0)
   • PASS if exits 0
   • Note: When run inside Claude Code, this test may show BLOCKED because the live hook intercepts the outer bash command. The

     __SKILL__=release

     mechanism works correctly - this is a test isolation artifact. For manual verification, run outside Claude Code.

4. Allow normal commands: Test that non-release commands pass through

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"tool_input": {"command": "npm install express"}}' | bash "$PLUGIN_PATH/hooks/block-manual-release.sh"
   echo "Exit code: $?"
   

   • Expected: Allowed (exit code 0)
   • PASS if exits 0

Part 3: PostToolUse Hook (post-edit-check.sh)

1. Anti-pattern Detection: Test detection of forbidden words

   cd /tmp/bluera-base-test
   echo 'const fallback = true;' >> index.js
   git add index.js
   echo '{"tool_name": "Edit", "tool_input": {"file_path": "index.js"}}' | \
     CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "${CLAUDE_PLUGIN_ROOT:-$(pwd)}/hooks/post-edit-check.sh" 2>&1
   EXIT=$?
   git checkout index.js 2>/dev/null
   echo "Exit code: $EXIT"
   

   • Expected: Detects "fallback" anti-pattern, exit code 2
   • PASS if output mentions anti-pattern and exits 2

2. Clean Code Passes: Test that clean code passes validation

   cd /tmp/bluera-base-test
   echo 'const clean = true;' >> index.js
   git add index.js
   echo '{"tool_name": "Edit", "tool_input": {"file_path": "index.js"}}' | \
     CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "${CLAUDE_PLUGIN_ROOT:-$(pwd)}/hooks/post-edit-check.sh" 2>&1
   EXIT=$?
   git checkout index.js 2>/dev/null
   echo "Exit code: $EXIT"
   

   • Expected: No output, exit code 0
   • PASS if exits 0

Part 4: Stop Hook (milhouse-stop.sh)

1. No State File: Test that hook exits cleanly when no milhouse loop active

   cd /tmp/bluera-base-test
   rm -rf .bluera/bluera-base/state/milhouse-loop.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"transcript_path": "/tmp/test.jsonl"}' | bash "$PLUGIN_PATH/hooks/milhouse-stop.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Silent exit 0 (no active loop)
   • PASS if exits 0 with no output

2. Invalid Iteration: Test handling of corrupted state file

   cd /tmp/bluera-base-test
   mkdir -p .bluera/bluera-base/state
   printf '%s\n' '---' 'iteration: invalid' 'max_iterations: 5' 'completion_promise: "done"' '---' '' 'test prompt' > .bluera/bluera-base/state/milhouse-loop.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"transcript_path": "/tmp/test.jsonl"}' | bash "$PLUGIN_PATH/hooks/milhouse-stop.sh" 2>&1
   EXIT=$?
   echo "Exit code: $EXIT"
   

   • Expected: Warning about invalid iteration, removes file, exits 0
   • PASS if mentions "invalid" and exits 0

Part 5: Milhouse Setup Hook

1. Setup Creates State File: Test milhouse-setup.sh creates proper state

   cd /tmp/bluera-base-test
   rm -rf .bluera/bluera-base/state/milhouse-loop.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/milhouse-setup.sh" --inline "Build the feature" --max-iterations 10 2>&1
   cat .bluera/bluera-base/state/milhouse-loop.md 2>/dev/null | head -10
   

   • Expected: State file created with iteration: 1, max_iterations: 10
   • PASS if file exists with correct fields

Part 6: Slash Commands (Invocation Test)

1. Commit Command Available: Run

   /bluera-base:commit

   (will show clean status)
   • Expected: Shows git status and workflow instructions
   • PASS if command executes without error

2. Cancel Milhouse Available: Verify cancel command works when no loop active

   rm -rf /tmp/bluera-base-test/.bluera/bluera-base/state/milhouse-loop.md
   

   Then run

   /bluera-base:cancel-milhouse

   • Expected: Message about no active loop
   • PASS if command executes

Part 7: Skills Verification

1. All Skills Have SKILL.md: Verify every skill directory contains SKILL.md

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   MISSING=0
   CHECKED=0
   for dir in "$PLUGIN_PATH/skills/"*/; do
     # Skip template-only directories (contain .template but no SKILL.md)
     if find "$dir" -maxdepth 1 -name "*.template" 2>/dev/null | grep -q . && [ ! -f "$dir/SKILL.md" ]; then
       continue
     fi
     CHECKED=$((CHECKED + 1))
     if [ ! -f "$dir/SKILL.md" ]; then
       echo "MISSING: $dir/SKILL.md"
       MISSING=$((MISSING + 1))
     fi
   done
   echo "Checked $CHECKED skill directories, $MISSING missing SKILL.md"
   [ $MISSING -eq 0 ] && echo "All skills valid"
   

   • Expected: All skill directories (except template-only) contain SKILL.md
   • PASS if output shows "All skills valid"

2. Skills Are Readable: Verify at least one skill file has content

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   # Find first skill with SKILL.md (skip template-only dirs)
   FIRST_SKILL=""
   for dir in "$PLUGIN_PATH/skills/"*/; do
     if [ -f "$dir/SKILL.md" ]; then
       FIRST_SKILL="$dir"
       break
     fi
   done
   [ -n "$FIRST_SKILL" ] && head -5 "$FIRST_SKILL/SKILL.md"
   

   • Expected: Shows skill header/title
   • PASS if file is readable with content

Part 8: Library Unit Tests

1. Signals Library Tests: Run signals.sh unit tests

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   bash "$PLUGIN_PATH/hooks/lib/tests/test-signals.sh"
   

   • Expected: 6 tests pass
   • PASS if output contains "6 passed, 0 failed"

2. State Library Tests: Run state.sh unit tests

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   bash "$PLUGIN_PATH/hooks/lib/tests/test-state.sh"
   

   • Expected: 8 tests pass
   • PASS if output contains "8 passed, 0 failed"

3. Gitignore Integration Tests: Run gitignore tests

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   bash "$PLUGIN_PATH/hooks/lib/tests/test-gitignore-integration.sh"
   

   • Expected: 4 tests pass
   • PASS if output contains "4 passed, 0 failed"

Part 9: Checklist Hook

1. Checklist Hook - No File: Test that hook exits cleanly when no checklist exists

   cd /tmp/bluera-base-test
   rm -rf .bluera/bluera-base/checklist.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/checklist-remind.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Silent exit 0 (no checklist file)
   • PASS if exits 0 with no output

2. Checklist Hook - With Items: Test that hook returns context when checklist has items

   cd /tmp/bluera-base-test
   mkdir -p .bluera/bluera-base
   printf '%s\n' '# Test Checklist' '' '[ ] Item 1' '[ ] Item 2' > .bluera/bluera-base/checklist.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/checklist-remind.sh" 2>&1
   EXIT=$?
   echo "Exit code: $EXIT"
   

   • Expected: JSON output with additionalContext containing "2 items pending"
   • PASS if output contains "CHECKLIST" and "2 items pending"

3. Checklist Hook - All Completed: Test that hook exits silently when all items checked

   cd /tmp/bluera-base-test
   printf '%s\n' '# Test Checklist' '' '[x] Item 1' '[x] Item 2' > .bluera/bluera-base/checklist.md
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/checklist-remind.sh" 2>&1
   EXIT=$?
   rm -rf .bluera/bluera-base/checklist.md
   echo "Exit code: $EXIT"
   

   • Expected: Silent exit 0 (no pending items)
   • PASS if exits 0 with no JSON output

Part 10: TODO Command

1. TODO File Creation: Test that TODO.txt can be created

   cd /tmp/bluera-base-test
   rm -rf .bluera/bluera-base/TODO.txt
   mkdir -p .bluera/bluera-base
   # Create minimal TODO file with required sections
   printf '%s\n' "## IMPORTANT" "" "* test item" "" "## TODO TASKS" "" "(none)" "" "## COMPLETED TASKS" "" "(none)" > .bluera/bluera-base/TODO.txt
   test -f .bluera/bluera-base/TODO.txt && echo "File created"
   

   • Expected: TODO.txt file is created
   • PASS if file exists

2. TODO File Structure: Verify TODO.txt has required sections

   cd /tmp/bluera-base-test
   grep -q "## TODO TASKS" .bluera/bluera-base/TODO.txt && \
   grep -q "## COMPLETED TASKS" .bluera/bluera-base/TODO.txt && \
   grep -q "## IMPORTANT" .bluera/bluera-base/TODO.txt && \
   echo "Structure valid"
   

   • Expected: All required sections present
   • PASS if all sections found

Part 11: Safe Slash Commands

Note: These commands are safe to run - they don't modify the repo structure.

1. /help: Run

   /bluera-base:help

   • Expected: Output contains "Commands" section
   • PASS if command executes and shows command list

2. /explain overview: Run

   /bluera-base:explain overview

   • Expected: Output contains "What is Bluera Base"
   • PASS if command executes and shows overview

3. /config show: Run

   /bluera-base:config show

   • Expected: Output shows config JSON or defaults
   • PASS if command executes without error

4. /checklist show: Run

   /bluera-base:checklist show

   • Expected: Shows checklist or "No checklist exists"
   • PASS if command executes without error

5. /todo show: Run

   /bluera-base:todo show

   • Expected: Shows TODO or creates file
   • PASS if command executes without error

6. /learn show: Run

   /bluera-base:learn show

   • Expected: Shows learnings or "No learnings"
   • PASS if command executes without error

7. /worktree list: Run

   /bluera-base:worktree list

   • Expected: Shows worktrees or "No worktrees"
   • PASS if command executes without error

8. /claude-code-statusline: Run

   /bluera-base:claude-code-statusline

   • Expected: Shows presets or current config
   • PASS if command executes without error

9. /claude-code-analyze-config: Run

   /bluera-base:claude-code-analyze-config

   • Expected: Shows analysis or "No .claude config"
   • PASS if command executes without error

10. /claude-code-md audit: Run

    /bluera-base:claude-code-md audit

    • Expected: Shows audit results
    • PASS if command executes without error

Part 12: Additional Hook Tests

1. notify.sh - Permission Prompt:

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"notification_type": "permission_prompt", "message": "Test"}' | \
     bash "$PLUGIN_PATH/hooks/notify.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (notification sent or no notifier available)
   • PASS if exits 0

2. session-setup.sh:

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/session-setup.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (jq check, permissions fix)
   • PASS if exits 0

3. session-start-inject.sh:

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/session-start-inject.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (may output context JSON)
   • PASS if exits 0

4. auto-commit.sh - No Changes:

   cd /tmp/bluera-base-test
   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{}' | CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/auto-commit.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (no uncommitted changes to prompt about)
   • PASS if exits 0

5. dry-scan.sh - Feature Disabled:

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{}' | CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/dry-scan.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (feature disabled by default)
   • PASS if exits 0

6. pre-compact.sh:

   PLUGIN_PATH="${CLAUDE_PLUGIN_ROOT:-$(pwd)}"
   echo '{"trigger": "manual"}' | CLAUDE_PROJECT_DIR="/tmp/bluera-base-test" bash "$PLUGIN_PATH/hooks/pre-compact.sh" 2>&1
   echo "Exit code: $?"
   

   • Expected: Exit 0 (preserves state before compaction)
   • PASS if exits 0

Part 13: Cleanup

1. Remove Test Directory: Clean up test artifacts

   rm -rf /tmp/bluera-base-test
   rm -rf .bluera/bluera-base/state/milhouse-loop.md
   

   • Expected: Directory removed
   • PASS if command succeeds

2. Verify Cleanup: Confirm test directory is gone

   ls /tmp/bluera-base-test 2>&1 || echo "Cleanup successful"
   

   • Expected: Directory not found
   • PASS if directory doesn't exist

Output Format

After running all tests, report results in this format:

Plugin Test Results

Result: X/41 tests passed

Error Recovery

If tests fail partway through, clean up manually:

rm -rf /tmp/bluera-base-test
rm -rf .bluera/bluera-base/state/milhouse-loop.md