# Test Plan: Tool Parsing Simplification

## Unit Tests

- [x] Test case 1: Single tool call → Returns 1 tool with correct name and arguments
- [x] Test case 2: No tool in text → Returns None for tools, original text as content  
- [x] Test case 3: Multiple tools → Returns all tools in order
- [x] Test case 4: Content before tool → Content extracted, tool parsed correctly
- [x] Test case 5: Bash tool → Correctly parses bash command
- [x] Test case 6: Case insensitive → "tool:" and "TOOL:" both work
- [x] Test case 7: Invalid JSON → Skips invalid, continues with valid
- [x] Test case 8: Empty text → Returns None, empty string
- [x] Test case 9: Whitespace only → Returns None

## Integration Tests

- [ ] End-to-end flow: 
  1. Send chat completion request with tools
  2. Model responds with TOOL:/ARGUMENTS: format
  3. Parser extracts tool call
  4. Tool executes
  5. Result returned in response

- [ ] Expected result: Tool executes successfully, result included in response

## Manual Verification

- [ ] Command: `python tests/test_tool_parsing.py`
- [ ] Expected output: "9 passed, 0 failed"

## Token Budget Verification

- Parser code: ~30 lines (~200 tokens)
- Well under 2000 token limit
- Simple regex pattern maintains low complexity