This is the meta issue about PinchBench v2 and what would be required to either (1) fix or (2) explicitly decide to not address to then make a new release and feel good about calling it a major release

Skill Changes

• Discussion: What tasks should we add next? #52
  • RFC: 20 real-world task ideas from ClawBytes use cases #124
  • Task ideas from WildClawBench analysis #121
  • Task ideas from canitrunopenclaw/ClawBench analysis #122
  • RFC: Additional task ideas - Email, Finance, Research, and More #125
• RFC: Mock services infrastructure for API-dependent tasks #123
• Fix or remove task that fails 95% of the time on 90% of models #51
• Feature: Publish per-task variance and retry counts on leaderboard #8
• benchmark: wait for transcript completion instead of assuming subprocess exit means task finished #65
• Thinking Levels #9
• Ensure that all "assets" are used in at least 1 task #178

Leaderboard Changes

• Redesign: Left sidebar with categorized filters, search input, and checkbox-style selections leaderboard#39
• Feature: User profiles and contributor leaderboard leaderboard#36
• Leaderboard scoring bias: runs with fewer tasks rank higher than comprehensive runs #53
• Daily, weekly, monthly badges leaderboard#48

Completed

• Discussion: Average vs. Best score leaderboard#42
• Feature: Model landing pages and search leaderboard#35
• Show overall rank of every run leaderboard#47
• feat: add task selection filter for benchmark runs leaderboard#22

Other

• Discussion: how should we handle different providers & hardware on the main results #47

Completed

• Meta: Migrate from Git Hash Versioning to Semantic Versioning #62