test: Add E2E test skeletons for 26 remaining autonomy scenarios

Created 7 test files covering all remaining scenarios for 100% autonomy:

High Priority (19 scenarios):
- complex_modifications_test.go (12): Multi-file refactoring, DB migrations, API changes, security fixes, etc.
- test_generation_test.go (5): Unit tests, integration tests, mocks, snapshots
- advanced_validation_test.go (2): Self-review, E2E test creation

Medium Priority (4 scenarios):
- codebase_understanding_test.go (2): Dependency tracing, doc parsing
- merge_conflicts_test.go (1): Merge conflict resolution

Low Priority (8 scenarios):
- advanced_git_test.go (5): Rebase, cherry-pick, bisect
- documentation_test.go (3): README, CHANGELOG, API docs

All tests initially marked with t.Skip() and descriptive TODO messages.
Tests compile and run successfully (30 skipped subtests).

Created permanent capabilities.md documentation replacing temporary
AUTONOMY_GAP_ANALYSIS.md. This doc will be updated with each release.

Removed obsolete analysis files:
- E2E_ANALYSIS.md, CAPABILITIES_ROADMAP.md, CAPABILITIES_TEST_RESULTS.md
- KNOWN_ISSUES.md, KNOWN_LIMITATIONS.md, QUICK_TEST_RESULTS.md
- TODO_refactor_model_search.md
- *.log files

This provides a clear roadmap: as we implement each capability, we can
enable the corresponding test to track progress toward 100% autonomy.

Author: jadercorrea