feat: rename README sections to CodeScaleBench-SDLC/Org, add build diary pipeline

- Rename "Benchmark Suites (SDLC-Aligned)" → "CodeScaleBench-SDLC"
- Rename "MCP-Unique Suites (Org-Scale Context Retrieval)" → "CodeScaleBench-Org"
- Add build diary extraction and visualization scripts:
  - extract_build_diary.py: parse 951 conversation transcripts into daily/session CSVs
  - extract_build_narrative.py: topic heatmap, commit arc, and rework data
  - plot_build_diary.py: 4-panel mechanical overview (activity, tools, ratio, distribution)
  - plot_build_narrative.py: 3-panel content figure (topics, arcs, rework curve)
  - plot_build_diary_supplementary.py: tool transitions, archetypes, repo hotspots
- Fix task ID extraction in convert_harbor_to_contextbench.py (job.log matching)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: sjarmak

README.md

@@ -64,7 +64,7 @@ bash configs/run_selected_tasks.sh --dry-run
 
 ---
 
-## Benchmark Suites (SDLC-Aligned)
+## CodeScaleBench-SDLC
 
 Nine suites organized by software development lifecycle phase:
 
@@ -81,7 +81,7 @@ Nine suites organized by software development lifecycle phase:
 | `csb_sdlc_debug` | Debugging & Investigation | 20 | Root cause tracing, fault localization, provenance |
 | **Total** | | **180** | |
 
-## MCP-Unique Suites (Org-Scale Context Retrieval)
+## CodeScaleBench-Org
 
 Eleven additional suites measure cross-repo discovery, symbol resolution, dependency tracing, and deep-search-driven investigation in polyrepo environments.