@hustle-together/api-dev-tools 3.10.1 → 3.11.1

package/.skills/_shared/hooks/enforce-interview.py

@@ -0,0 +1,421 @@
+#!/usr/bin/env python3
+"""
+Hook: PreToolUse for Write/Edit
+Purpose: Block proceeding to schema/TDD if interview has no USER answers
+
+This hook ensures Claude actually asks the user questions and records
+their answers, rather than self-answering the interview.
+
+v1.8.0 MAJOR UPDATE: Now requires STRUCTURED questions with multiple-choice
+options derived from research phase findings.
+
+It checks:
+  1. Research phase is complete (questions must be based on research)
+  2. Interview status is "complete"
+  3. Questions used AskUserQuestion tool with STRUCTURED OPTIONS
+  4. At least MIN_STRUCTURED_QUESTIONS have multiple-choice or typed options
+  5. Answers don't look auto-generated (contain user-specific details)
+
+The goal: Questions like Claude Code shows - with numbered options and
+"Type something" at the end, all based on research findings.
+
+Returns:
+  - {"permissionDecision": "allow"} - Let the tool run
+  - {"permissionDecision": "deny", "reason": "..."} - Block with explanation
+"""
+import json
+import sys
+from pathlib import Path
+
+# State file is in .claude/ directory (sibling to hooks/)
+STATE_FILE = Path(__file__).parent.parent / "api-dev-state.json"
+
+# Minimum questions required for a valid interview
+MIN_QUESTIONS = 5  # Increased - need comprehensive interview
+
+# Minimum questions that MUST have structured options (multiple-choice)
+MIN_STRUCTURED_QUESTIONS = 3
+
+# Phrases that indicate self-answered (not real user input)
+SELF_ANSWER_INDICATORS = [
+    "based on common",
+    "self-answered",
+    "assumed",
+    "typical use case",
+    "standard implementation",
+    "common pattern",
+    "i'll assume",
+    "assuming",
+    "probably",
+    "most likely",
+    "default to",
+    "usually",
+]
+
+
+def main():
+    # Read hook input from stdin
+    try:
+        input_data = json.load(sys.stdin)
+    except json.JSONDecodeError:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    tool_input = input_data.get("tool_input", {})
+    file_path = tool_input.get("file_path", "")
+
+    # Enforce for ANY file in /api/ directory (not just route.ts)
+    # This forces Claude to stop and interview before ANY API work
+    is_api_file = "/api/" in file_path and file_path.endswith(".ts")
+    is_schema_file = "/schemas/" in file_path and file_path.endswith(".ts")
+
+    # Skip test files - those are allowed during TDD
+    is_test_file = ".test." in file_path or "/__tests__/" in file_path or ".spec." in file_path
+
+    if is_test_file:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    if not is_schema_file and not is_api_file:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Check if state file exists
+    if not STATE_FILE.exists():
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": """❌ API workflow not started.
+
+Run /api-create [endpoint-name] to begin the interview-driven workflow."""
+        }))
+        sys.exit(0)
+
+    # Load state
+    try:
+        state = json.loads(STATE_FILE.read_text())
+    except json.JSONDecodeError:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    phases = state.get("phases", {})
+    research = phases.get("research_initial", {})
+    interview = phases.get("interview", {})
+    interview_status = interview.get("status", "not_started")
+    interview_desc = interview.get("description", "").lower()
+    questions = interview.get("questions", [])
+    research_queries = state.get("research_queries", [])
+
+    # Check 0: Research must be complete FIRST (questions based on research)
+    research_status = research.get("status", "not_started")
+    if research_status != "complete":
+        sources_count = len(research.get("sources", []))
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ BLOCKED: Research phase must complete BEFORE interview.
+
+Research status: {research_status}
+Sources consulted: {sources_count}
+Research queries: {len(research_queries)}
+
+═══════════════════════════════════════════════════════════
+⚠️  COMPLETE RESEARCH FIRST - THEN ASK QUESTIONS
+═══════════════════════════════════════════════════════════
+
+The interview questions MUST be based on research findings:
+1. Use Context7 to get SDK/API documentation
+2. Use WebSearch (2-3 searches) for official docs
+3. THEN generate interview questions with STRUCTURED OPTIONS
+   based on what you discovered
+
+Example: If research found 5 available models, ask:
+  "Which model should this endpoint use?"
+  1. gpt-4o (fastest, cheapest)
+  2. claude-sonnet-4-20250514 (best reasoning)
+  3. gemini-pro (multimodal)
+  4. Type something else...
+
+Research INFORMS the options. No research = no good options."""
+        }))
+        sys.exit(0)
+
+    # Check 1: Interview must be complete
+    if interview_status != "complete":
+        # Build example based on actual research
+        research_based_example = _build_research_based_example(research_queries)
+
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ BLOCKED: Interview phase not complete.
+
+Current status: {interview_status}
+AskUserQuestion calls: {interview.get('user_question_count', 0)}
+Structured questions: {interview.get('structured_question_count', 0)}
+
+═══════════════════════════════════════════════════════════
+⚠️  USE STRUCTURED QUESTIONS WITH OPTIONS
+═══════════════════════════════════════════════════════════
+
+Based on your research, ask questions using AskUserQuestion with
+the 'options' parameter to provide multiple-choice selections:
+
+{research_based_example}
+
+REQUIRED FORMAT for AskUserQuestion:
+- question: "Your question text"
+- options: [
+    {{"value": "option1", "label": "Option 1 description"}},
+    {{"value": "option2", "label": "Option 2 description"}},
+    {{"value": "custom", "label": "Type something..."}}
+  ]
+
+You need at least {MIN_STRUCTURED_QUESTIONS} structured questions with options.
+Current: {interview.get('structured_question_count', 0)}
+
+DO NOT:
+❌ Ask open-ended questions without options
+❌ Make up options not based on research
+❌ Skip the AskUserQuestion tool
+❌ Self-answer questions"""
+        }))
+        sys.exit(0)
+
+    # Check 2: Must have minimum questions
+    if len(questions) < MIN_QUESTIONS:
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ Interview incomplete - not enough questions asked.
+
+Questions recorded: {len(questions)}
+Minimum required: {MIN_QUESTIONS}
+
+You must ask the user more questions about their requirements.
+Use AskUserQuestion with structured options based on your research."""
+        }))
+        sys.exit(0)
+
+    # Check 3: Verify AskUserQuestion tool was actually used
+    user_question_count = interview.get("user_question_count", 0)
+    tool_used_count = sum(1 for q in questions if q.get("tool_used", False))
+
+    if tool_used_count < MIN_QUESTIONS:
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ Interview not conducted properly.
+
+AskUserQuestion tool uses tracked: {tool_used_count}
+Minimum required: {MIN_QUESTIONS}
+
+You MUST use the AskUserQuestion tool to ask the user directly.
+Do NOT make up answers or mark the interview as complete without
+actually asking the user and receiving their responses."""
+        }))
+        sys.exit(0)
+
+    # Check 4: Verify structured questions were used
+    structured_count = interview.get("structured_question_count", 0)
+    questions_with_options = sum(1 for q in questions if q.get("has_options", False))
+    actual_structured = max(structured_count, questions_with_options)
+
+    if actual_structured < MIN_STRUCTURED_QUESTIONS:
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ Not enough STRUCTURED questions with options.
+
+Structured questions (with options): {actual_structured}
+Minimum required: {MIN_STRUCTURED_QUESTIONS}
+
+You MUST use AskUserQuestion with the 'options' parameter to
+provide multiple-choice answers based on your research.
+
+Example:
+  AskUserQuestion(
+    question="Which AI provider should this endpoint support?",
+    options=[
+      {{"value": "openai", "label": "OpenAI (GPT-4o)"}},
+      {{"value": "anthropic", "label": "Anthropic (Claude)"}},
+      {{"value": "google", "label": "Google (Gemini)"}},
+      {{"value": "all", "label": "All of the above"}},
+      {{"value": "custom", "label": "Type something else..."}}
+    ]
+  )
+
+This gives the user clear choices based on what you researched."""
+        }))
+        sys.exit(0)
+
+    # Check 5: Look for self-answer indicators
+    for indicator in SELF_ANSWER_INDICATORS:
+        if indicator in interview_desc:
+            print(json.dumps({
+                "permissionDecision": "deny",
+                "reason": f"""❌ Interview appears to be self-answered.
+
+Detected: "{indicator}" in interview description.
+
+You MUST actually ask the user questions using AskUserQuestion
+with structured options. Self-answering defeats the purpose.
+
+Reset the interview and ask with options based on research."""
+            }))
+            sys.exit(0)
+
+    # Check 6: FINAL USER CONFIRMATION - must confirm interview is complete
+    user_question_asked_final = interview.get("user_question_asked", False)
+    user_completed = interview.get("user_completed", False)
+    phase_exit_confirmed = interview.get("phase_exit_confirmed", False)
+    decisions = interview.get("decisions", {})
+
+    if not user_completed or not user_question_asked_final or not phase_exit_confirmed:
+        decision_summary = _build_decision_summary(decisions)
+        missing = []
+        if not user_question_asked_final:
+            missing.append("Final confirmation question (AskUserQuestion not used)")
+        if not user_completed:
+            missing.append("User hasn't confirmed interview complete")
+        if not phase_exit_confirmed:
+            missing.append("Phase exit confirmation (user must explicitly approve to proceed)")
+
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ BLOCKED: Interview needs FINAL USER CONFIRMATION.
+
+Questions asked: {len(questions)}
+Structured questions: {actual_structured}
+User final confirmation: {user_completed}
+Phase exit confirmed: {phase_exit_confirmed}
+
+MISSING:
+{chr(10).join(f"  • {m}" for m in missing)}
+
+═══════════════════════════════════════════════════════════
+⚠️  GET USER CONFIRMATION BEFORE PROCEEDING
+═══════════════════════════════════════════════════════════
+
+REQUIRED STEPS:
+
+1. SHOW interview summary to user:
+   ┌───────────────────────────────────────────────────────┐
+   │ INTERVIEW COMPLETE                                    │
+   │                                                       │
+   │ Your decisions:                                       │
+{chr(10).join(f"   │   • {line:<49} │" for line in decision_summary.split(chr(10))[:8]) if decision_summary else "   │   (no decisions recorded yet)                      │"}
+   │                                                       │
+   │ These will guide the schema, tests, and implementation│
+   │                                                       │
+   │ All correct? [Y]                                      │
+   │ Change an answer? [n] ____                            │
+   └───────────────────────────────────────────────────────┘
+
+2. USE AskUserQuestion:
+   question: "Interview decisions correct? Ready to proceed?"
+   options: [
+     {{"value": "confirm", "label": "Yes, proceed to schema creation"}},
+     {{"value": "change", "label": "No, I want to change [which question]"}},
+     {{"value": "add", "label": "Add another question about [topic]"}}
+   ]
+
+3. If user says "change" or "add":
+   • Ask which question/topic
+   • Re-ask with AskUserQuestion
+   • Update decisions
+   • LOOP BACK and show updated summary
+
+4. If user says "confirm":
+   • Set interview.user_question_asked = true
+   • Set interview.user_completed = true
+   • Set interview.status = "complete"
+
+WHY: User must approve their decisions before they drive implementation."""
+        }))
+        sys.exit(0)
+
+    if decisions:
+        # Build a reminder of what the user decided
+        decision_summary = _build_decision_summary(decisions)
+
+        # Allow but inject context about user decisions
+        print(json.dumps({
+            "permissionDecision": "allow",
+            "message": f"""✅ Interview complete. REMEMBER THE USER'S DECISIONS:
+
+{decision_summary}
+
+Your implementation MUST align with these choices.
+The state file tracks these for consistency verification."""
+        }))
+    else:
+        print(json.dumps({"permissionDecision": "allow"}))
+
+    sys.exit(0)
+
+
+def _build_decision_summary(decisions: dict) -> str:
+    """Build a human-readable summary of user decisions from the interview."""
+    if not decisions:
+        return "No key decisions recorded."
+
+    lines = []
+    decision_labels = {
+        "provider": "AI Provider",
+        "purpose": "Primary Purpose",
+        "response_format": "Response Format",
+        "required_params": "Required Parameters",
+        "optional_params": "Optional Parameters",
+        "error_handling": "Error Handling",
+        "api_key_handling": "API Key Handling",
+        "external_services": "External Services",
+    }
+
+    for key, data in decisions.items():
+        label = decision_labels.get(key, key.replace("_", " ").title())
+        response = data.get("response", "")
+        value = data.get("value", "")
+
+        if value:
+            lines.append(f"• {label}: {value}")
+        elif response:
+            # Truncate long responses
+            short_response = response[:80] + "..." if len(response) > 80 else response
+            lines.append(f"• {label}: {short_response}")
+
+    return "\n".join(lines) if lines else "No key decisions recorded."
+
+
+def _build_research_based_example(research_queries: list) -> str:
+    """Build an example question based on actual research queries."""
+    if not research_queries:
+        return """Example (generic - do research first!):
+  "What is the main use case for this endpoint?"
+  1. Data retrieval
+  2. Data transformation
+  3. AI processing
+  4. Type something..."""
+
+    # Extract terms from research to suggest relevant options
+    all_terms = []
+    for query in research_queries[-5:]:  # Last 5 queries
+        terms = query.get("terms", [])
+        all_terms.extend(terms)
+
+    # Deduplicate and get top terms
+    unique_terms = list(dict.fromkeys(all_terms))[:4]
+
+    if unique_terms:
+        options_example = "\n  ".join([
+            f"{i+1}. {term.title()}" for i, term in enumerate(unique_terms)
+        ])
+        return f"""Example based on your research:
+  "Which of these should be the primary focus?"
+  {options_example}
+  {len(unique_terms)+1}. Type something else..."""
+
+    return """Example:
+  "What capability is most important?"
+  1. Option based on research finding 1
+  2. Option based on research finding 2
+  3. Option based on research finding 3
+  4. Type something..."""
+
+
+if __name__ == "__main__":
+    main()

package/.skills/_shared/hooks/enforce-refactor.py

@@ -0,0 +1,189 @@
+#!/usr/bin/env python3
+"""
+Hook: PreToolUse for Write/Edit
+Purpose: Block refactoring until verification phase is complete
+
+This hook ensures that after tests pass (Green phase), the implementation
+is verified against documentation before any refactoring begins.
+
+Phase 20 of the 13-phase workflow requires:
+  - TDD Green phase complete (tests passing)
+  - Verify phase complete (Phase 10)
+  - All gaps found have been fixed or documented as intentional omissions
+
+Returns:
+  - {"permissionDecision": "allow"} - Let the tool run
+  - {"permissionDecision": "deny", "reason": "..."} - Block with explanation
+"""
+import json
+import sys
+from pathlib import Path
+
+# State file is in .claude/ directory (sibling to hooks/)
+STATE_FILE = Path(__file__).parent.parent / "api-dev-state.json"
+
+# Keywords that suggest refactoring intent
+REFACTOR_KEYWORDS = [
+    "refactor",
+    "cleanup",
+    "clean up",
+    "restructure",
+    "reorganize",
+    "optimize",
+    "simplify",
+    "extract",
+    "rename",
+    "move",
+]
+
+
+def is_refactoring_edit(tool_input: dict) -> bool:
+    """Detect if this edit appears to be a refactoring operation."""
+    # Check the content being written
+    new_string = tool_input.get("new_string", "")
+    old_string = tool_input.get("old_string", "")
+
+    # If both old and new exist and are similar length, might be refactoring
+    if old_string and new_string:
+        len_diff = abs(len(new_string) - len(old_string))
+        # Similar length suggests refactoring vs new features
+        if len_diff < len(old_string) * 0.3:  # Within 30% length
+            return True
+
+    return False
+
+
+def main():
+    # Read hook input from stdin
+    try:
+        input_data = json.load(sys.stdin)
+    except json.JSONDecodeError:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    tool_name = input_data.get("tool_name", "")
+    tool_input = input_data.get("tool_input", {})
+    file_path = tool_input.get("file_path", "")
+
+    # Only check Edit operations on API files (Write is for new content)
+    if tool_name != "Edit":
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Only enforce for API route files
+    if "/api/" not in file_path or not file_path.endswith(".ts"):
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Skip test files - can refactor tests anytime
+    if ".test." in file_path or "/__tests__/" in file_path or ".spec." in file_path:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Check if state file exists
+    if not STATE_FILE.exists():
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Load state
+    try:
+        state = json.loads(STATE_FILE.read_text())
+    except json.JSONDecodeError:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    endpoint = state.get("endpoint")
+    if not endpoint:
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    phases = state.get("phases", {})
+    tdd_green = phases.get("tdd_green", {})
+    verify = phases.get("verify", {})
+    tdd_refactor = phases.get("tdd_refactor", {})
+
+    # Only enforce after TDD Green is complete
+    if tdd_green.get("status") != "complete":
+        # Still in implementation phase, allow edits
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Check if this looks like a refactoring edit
+    if not is_refactoring_edit(tool_input):
+        # Doesn't look like refactoring, might be bug fix
+        print(json.dumps({"permissionDecision": "allow"}))
+        sys.exit(0)
+
+    # Check verify phase status
+    verify_status = verify.get("status", "not_started")
+    gaps_found = verify.get("gaps_found", 0)
+    gaps_fixed = verify.get("gaps_fixed", 0)
+    intentional_omissions = verify.get("intentional_omissions", [])
+    phase_exit_confirmed = verify.get("phase_exit_confirmed", False)
+
+    if verify_status != "complete" or not phase_exit_confirmed:
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ BLOCKED: Verify phase (Phase 10) not complete.
+
+Current status: {verify_status}
+Gaps found: {gaps_found}
+Gaps fixed: {gaps_fixed}
+Intentional omissions: {len(intentional_omissions)}
+Phase exit confirmed: {phase_exit_confirmed}
+
+═══════════════════════════════════════════════════════════
+⚠️  VERIFY BEFORE REFACTORING
+═══════════════════════════════════════════════════════════
+
+Before refactoring, you must:
+
+1. Re-read the original documentation
+2. Compare implementation to docs feature-by-feature
+3. Fix any gaps OR document them as intentional omissions
+
+Current gaps not addressed:
+  • {gaps_found - gaps_fixed} gaps still need attention
+
+Once verify phase is complete:
+  • All gaps fixed OR documented as omissions
+  • Implementation matches documented behavior
+  • THEN you can safely refactor
+
+WHY THIS MATTERS:
+  - Refactoring should not change behavior
+  - Must verify behavior is CORRECT before preserving it
+  - Otherwise you cement bugs into clean code"""
+        }))
+        sys.exit(0)
+
+    # Verify complete - check if all gaps addressed
+    unaddressed_gaps = gaps_found - gaps_fixed - len(intentional_omissions)
+    if unaddressed_gaps > 0:
+        print(json.dumps({
+            "permissionDecision": "deny",
+            "reason": f"""❌ BLOCKED: {unaddressed_gaps} gaps not addressed.
+
+Gaps found: {gaps_found}
+Gaps fixed: {gaps_fixed}
+Intentional omissions: {len(intentional_omissions)}
+Unaddressed: {unaddressed_gaps}
+
+Fix the remaining gaps or mark them as intentional omissions
+before refactoring."""
+        }))
+        sys.exit(0)
+
+    # All clear for refactoring
+    refactor_status = tdd_refactor.get("status", "not_started")
+    print(json.dumps({
+        "permissionDecision": "allow",
+        "message": f"""✅ Verification complete. Safe to refactor.
+Refactor phase status: {refactor_status}
+Remember: Tests must still pass after refactoring!"""
+    }))
+    sys.exit(0)
+
+
+if __name__ == "__main__":
+    main()