fix(git): normalize git_log and git_show formatting across filtered and unfiltered queries

[teddiesloco]

Summary

Fixes the schema and formatting asymmetry between filtered (start_timestamp / end_timestamp) and unfiltered calls in git_log, as well as git_show.

Root Cause

1. Repr vs Plain String: In unfiltered git_log and git_show, commit.hexsha!r and commit.message!r were used, producing quoted strings like Commit: 'b9fd86...' and Message: 'feat: ...'. In contrast, the timestamp-filtered branch produced unquoted Commit: b9fd86....
2. Author Format Asymmetry: Filtered queries produced Author: Name <email> or plain author name, whereas unfiltered queries produced Author: <git.Actor "Name <email>">.
3. Date Format Asymmetry: Filtered queries used git CLI's human-readable format while unfiltered queries used Python datetime object string.
4. Delimiter Safety: Date-filtered git log now uses ASCII Unit/Record Separators (%x1e / %x1f) instead of splitting by raw newlines (\n), preventing multi-line commit messages from corrupting the parsed log entry count.

Fix

• Standardized both code paths to output clean unquoted Commit: <sha>, Author: Name <email>, ISO-8601 Date: <iso_datetime>, and clean Message: <message>.
• Added Unit Separator delimiters in git log command to handle arbitrary multi-line commit subjects safely.
• Added regression test test_git_log_schema_consistency asserting schema uniformity across filtered and unfiltered queries.

Verification

• Ran full test suite via uv run pytest tests/ -q (43/43 PASS).
• Verified monorepo build (npm run build clean across all workspace packages).