test: Complete Testing and Quality Assurance for v0.3

.claude/Guidelines.md

@@ -0,0 +1,91 @@
+# Gadugi Development Guidelines
+
+## CRITICAL: Zero BS Principle
+
+**NO BULLSHIT. NO CLAIMS WITHOUT EVIDENCE. NO FAKE COMPLETIONS.**
+
+- If code doesn't exist, say "NOT IMPLEMENTED"
+- If it's a stub, say "STUB ONLY"
+- If it's untested, say "UNTESTED"
+- If it doesn't work, say "BROKEN"
+- NEVER claim something is complete unless it actually works end-to-end
+
+## Core Development Principles
+
+### 1. Ruthless Honesty
+- Admit what's not done
+- Acknowledge what's broken
+- Report actual status, not aspirational status
+- If you haven't tested it, don't claim it works
+
+### 2. Implementation Before Claims
+- Write the code first
+- Test it second
+- Document it third
+- Claim completion only after all three
+
+### 3. Recipe-Driven Development
+Every component needs:
+- **Requirements**: What it MUST do (not what we hope it does)
+- **Design**: How it will actually work (not hand-waving)
+- **Implementation**: Real code that runs (not stubs)
+- **Tests**: Proof that it works (not hope)
+
+### 4. Quality Gates (MANDATORY)
+Before ANY code is considered complete:
+- ✅ Passes `uv run pyright` with ZERO errors
+- ✅ Formatted with `uv run ruff format`
+- ✅ Passes `uv run ruff check`
+- ✅ Has actual tests that pass with `uv run pytest`
+- ✅ Pre-commit hooks pass
+- ✅ Code review completed
+- ✅ System design review completed
+
+### 5. Dependency-Driven Order
+- Build foundations first
+- Don't build on top of stubs
+- Test each layer before building the next
+- If a dependency is broken, stop and fix it
+
+### 6. Testing Requirements
+- Every function needs a test
+- Every API endpoint needs integration tests
+- Every service needs end-to-end tests
+- No "it should work" - prove it works
+
+### 7. Review Requirements
+EVERY implementation needs:
+1. Design review (before coding)
+2. Code review (after coding)
+3. System design review (after integration)
+4. Sign-off from review agent
+
+## Implementation Checklist
+
+For EVERY component:
+- [ ] Recipe exists (requirements.md, design.md, dependencies.json)
+- [ ] Implementation matches recipe requirements
+- [ ] All dependencies are actually implemented (not stubs)
+- [ ] Unit tests exist and pass
+- [ ] Integration tests exist and pass
+- [ ] Pyright passes with zero errors
+- [ ] Ruff format and check pass
+- [ ] Pre-commit hooks configured and pass
+- [ ] Code review completed
+- [ ] System design review completed
+- [ ] Actually works when run (not just compiles)
+
+## Humility Principle
+- No performance claims without benchmarks
+- No "production-ready" claims without production testing
+- No "complete" claims without end-to-end validation
+- Let the code speak for itself
+
+## The Truth Test
+Before claiming anything:
+1. Can I run it right now?
+2. Does it actually do what the requirements say?
+3. Have I tested it with real data?
+4. Would I bet money that it works?
+
+If any answer is "no", then it's NOT DONE.

.claude/agent-manager/tests/test_checksum_verification.py

@@ -8,6 +8,7 @@
 import tempfile
 import unittest
 from pathlib import Path
+from typing import Set
 
 
 class TestChecksumVerification(unittest.TestCase):

.claude/agent-manager/tests/test_hook_setup.py

@@ -10,10 +10,9 @@
 import os
 import shutil
 import subprocess
-import sys
-import tempfile
 import unittest
 from pathlib import Path
+from typing import Set
 
 
 class TestAgentManagerHookSetup(unittest.TestCase):
@@ -173,7 +172,7 @@ def test_invalid_json_handling(self):
         with open(self.settings_file, 'w') as f:
             f.write('{"invalid": json content}')
 
-        result = self.run_setup_script()
+        _result = self.run_setup_script()
 
         # Should still create valid settings
         self.assertTrue(self.settings_file.exists())

.claude/agent-manager/tests/test_structure.py

@@ -7,6 +7,7 @@
 
 import unittest
 from pathlib import Path
+from typing import Set
 
 
 class TestAgentManagerStructure(unittest.TestCase):

.claude/agents/agent-updater.md

@@ -1,5 +1,6 @@
 ---
 name: agent-updater
+model: inherit
 description: Automatically checks for and manages updates for Claude Code agents, ensuring all agents are up-to-date
 tools: Read, Write, Edit, Bash, Grep, LS, TodoWrite, WebFetch
 ---

.claude/agents/code-review-response.md

@@ -1,5 +1,6 @@
 ---
 name: code-review-response
+model: inherit
 description: Processes code review feedback systematically, implements appropriate changes, and maintains professional dialogue throughout the review process
 tools: Read, Edit, MultiEdit, Bash, Grep, LS, TodoWrite
 ---

.claude/agents/code-reviewer.md

@@ -1,5 +1,6 @@
 ---
 name: code-reviewer
+model: inherit
 description: Specialized sub-agent for conducting thorough code reviews on pull requests
 tools: Read, Grep, LS, Bash, WebSearch, WebFetch, TodoWrite
 ---