- Refactored extractCommandPatterns to use the existing parseCommand function
- Ensures consistent command parsing behavior across the codebase
- Maintains security by removing subshell contents before parsing
- All existing tests continue to pass
- Fixed CommandExecution.tsx to only extract patterns from actual commands, not AI suggestions
- Enhanced extractCommandPatterns to filter out numeric patterns and common output words
- Added comprehensive test coverage for the bug scenario
- Ensures 'Manage Command Permissions' only shows actual executed commands
Fixes the issue where output like '0 total' from wc commands was incorrectly shown as a command pattern
- Fixed parseCommandAndOutput to properly handle the newline + 'Output:' separator
- Added test cases for commands with numbers at the start of output lines
- Updated existing tests to use template literals for proper newline handling
- Fixed test assertions to handle multiple code blocks when output is present
This resolves the issue where output lines starting with numbers (like 'wc -l' output)
were being incorrectly parsed as the command instead of the actual command text.
- Refactored showSuggestions from state variable to constant SHOW_SUGGESTIONS
- Renamed breakingExps to stopPatterns for better clarity
- Added test coverage for edge cases in command parsing
- Fixed test assertion for multiline content handling
