feat: add /debug-test slash command for analyzing test failures (#855)

Signed-off-by: emmaaroche <eroche@redhat.com>

Author: emmaaroche

.claude/commands/debug-test.md

@@ -0,0 +1,69 @@
+---
+description: Analyze test failures and suggest fixes
+---
+
+# Debug Test Failures
+
+Analyze test failures and suggest fixes.
+
+## Instructions
+
+Follow these steps to analyze the test failure:
+
+1. **Extract file path from traceback**: The error shows the failing file path (e.g., `testsuite/tests/.../test_basic.py:39`)
+
+2. **Read test file and trace fixtures**:
+   - Read the test file at the path from the traceback
+   - Look at fixtures (function parameters) and imports that the test uses
+   - Read `conftest.py` in same directory, then parent directories if needed
+   - Understand what the test expects vs what's actually happening
+
+3. **Categorize the failure**:
+   - Identify the pattern (see `Common Failure Patterns` below)
+   - Determine if this is a test/code issue or an environment/cluster state issue
+
+4. **Identify file authors**:
+   - For the failing test file and relevant conftest.py files, use this command to find recent contributors:
+     ```bash
+     git log --format='%an' <file> | sort | uniq -c | sort -rn | head -5
+     ```
+   - Include these authors in the output so users know who to contact for deeper context if needed
+
+5. **Provide appropriate response**:
+   - **If investigation needed**: Suggest kubectl commands to check policy status, conditions, events, logs, and verify resources exist and are ready
+     - **IMPORTANT**: kubectl commands often require resources still in the cluster. Since tests cleanup after finishing, instruct to:
+       - Set a breakpoint (outline where to set this)
+       - Re-run the failing test locally
+   - **If cluster/environment issue**: Provide kubectl/cleanup commands to fix cluster state
+   - **If test/code issue**: Suggest code changes to fixtures or tests
+   - **Mixed approach**: Combine multiple approaches as needed (e.g., investigate first, then apply fixes based on what you find)
+
+## Common Failure Patterns
+
+- **Policy not enforced**: Expected 302/401/429 but got 200 (policy not attached, wrong route/gateway, wrong host, or enforcement not yet propagated)
+- **Policy targeting**: `is_affected_by()` returns `False` (wrong `targetRef`, `parentRef`, namespace, labels, or resource mismatch)
+- **Reconciliation / propagation delay**: Async reconciliation not finished, config/policy not pushed, `observedGeneration` not catching up, race conditions
+- **DNS resolution / propagation**: `NXDOMAIN`, stale records (external DNS delay or caching, record deletion not fully propagated)
+- **DNS provider / credentials**: `DNSProviderError` (missing/misnamed/wrong-namespace secret, provider throttling or rate limits)
+- **Dependency missing or not found**: Policy `Accepted=False`, `reason=MissingDependency`, or selector found no objects (`selected 0`)
+  (required operator/backend not installed, wrong namespace, or wrong labels, e.g., Authorino, Limitador)
+- **Component not ready**: Dependency installed but not `Ready` (slow rollout or readiness probe delays)
+- **Connectivity / request timeouts**: Gateway or Listener not ready, LoadBalancer provisioning delay, DNS resolves but traffic does not route, client `httpx` connect/read timeouts
+- **Wrong endpoint hit**: Test traffic bypasses the intended Gateway or Route (wrong URL, scheme, port, or hostname)
+- **Stale state / cleanup issues**: Leftover resources from previous tests (policies, routes, DNS records, namespaces not fully deleted)
+- **Invalid or rejected configuration**: Policy `Accepted=False` with reasons like `Invalid` or `TargetNotFound` (bad references, missing targets, wrong issuer/secret)
+- **Status / model mismatch**: `MissingModelBranch` (testsuite expects status/conditions not present, CRD or controller output drift)
+
+## Output Format
+
+**Issue**: [Summary of the problem]
+
+**Root cause**: [Summary of why it's failing, or "needs more investigation" if unclear]
+
+**Recommended steps**: [One or more steps - can be investigation, code changes, environment fixes, or a mix]
+
+**File authors (Contact these contributors if you need deeper context beyond Claude's analysis)**:
+
+- `<file_path>`: Author Name (N commits), Another Author (M commits), ...
+
+_NOTE:_ Do NOT auto-run bash/kubectl commands or web searches before providing direct analysis and recommendations.