CRUX-in-a-box: open-source infra for running weeks-long agent evaluations
sayashk · x · 2026-11-06
Sayashk announces CRUX-in-a-box, open-source infrastructure distilled from six months of long-horizon agent evaluations (like having agents build and ship a mobile app). It deploys an AWS orchestrator via Agent Client Protocol, supports Claude Code and Codex on remote servers, provisions service access (GitHub, email, cloud compute), uses LLMs to flag risky actions for human review, and lets you inspect tool calls, tokens and spend with time and cost budgets.
Related event: CRUX-in-a-box Open-Sources Infrastructure for Weeks-Long Agent Evaluations(3 posts)→
More from coding & agent
- Alchemy teases next-week release: Rivet and Effect Cluster as swappable deploy layers — samgoodwin89 · 2026-11-06
- Gauge open-sources CLI as 85% of coding sessions are now agent-initiated — ycombinator · 2026-11-06
- Compound Security: open-source skill for agent-driven vulnerability patching — hamostaf04 · 2026-11-06
- Developer drives software factories and cloud agents via Slack, rarely opens a terminal — vikvang1 · 2026-11-06
- Glasshand rents real iPhones and Macs so AI agents can tap, type and swipe like humans — jia_seed · 2026-11-06
- No retraining needed: an LLM agent fixes a TTS model by editing weights directly — N1kuson · 2026-11-06