trajrl-bench 是做什麼的？
TrajRL-Bench: AI agent skills benchmark. SSH sandbox with mock services, LLM judge scoring, split-half delta evaluation. Leaderboard at trajrl.com/bench
TrajRL-Bench: AI agent skills benchmark. SSH sandbox with mock services, LLM judge scoring, split-half delta evaluation. Leaderboard at trajrl.com/bench
trajrl-bench is a Claude Code agent skill that trajRL-Bench: AI agent skills benchmark. SSH sandbox with mock services, LLM judge scoring, split-half delta evaluation. Leaderboard at trajrl.com/bench.
npx skills add trajectoryRL/trajrl-benchInstalled? Explore more 程式設計與開發 skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →
TrajRL-Bench: AI agent skills benchmark. SSH sandbox with mock services, LLM judge scoring, split-half delta evaluation. Leaderboard at trajrl.com/bench
Send and manage iMessages via BlueBubbles, including attachments, tapbacks, edits, replies, and groups.
Control Eight Sleep pods (status, temperature, alarms, schedules).
BluOS CLI (blu) for discovery, playback, grouping, and volume.
Create, search, and manage Bear notes via grizzly CLI.
Capture frames or clips from RTSP/ONVIF cameras.
Search GIF providers with CLI/TUI, download results, and extract stills/sheets.