[knight22-21]

Summary

Expands benchmark repository definitions and task suites to cover 11 corpora across 220 tasks:

• Registers 7 new repositories in \�enchmarks/setup_repos.py\ (\astapi, \pydantic, \express,
  eact, \ ypescript, \gin, \cobra).
• Adds dedicated task definition files for each new corpus (20 tasks each) and expands existing task suites (\ixture,
  equests, \lask, \httpx) to 20 tasks each.
• Updates \�enchmarks/RESULTS.md\ with Run 11 and Run 12 benchmark records across all 11 corpora.
• Updates \docs/benchmark-results.md\ summary table with the latest 220-task level 1 token reduction and accuracy results.

Testing

• Ran \python -m benchmarks.run_token_benchmark\ (20 fixture tasks completed successfully with +10.3% token reduction).
• Validated JSON syntax across all task definition files in \�enchmarks/tasks/.
• Verified
  uff check\ and
  uff format --check.