Long builds no longer die at the shell timeout, and compaction no longer stalls the next prompt. GPT 6.2 Sol and Astra are selectable with ChatGPT sign-in.
Key takeaways · AI-distilled
In OpenCode 2.3.1, a shell command still running when its timeout expires moves to the background instead of being killed, so long builds and test runs finish and report back. A foreground `timeout: 0` now means the default 120 seconds.
Background context compactionSummarizing an agent's earlier conversation to free room in the context window so a long session can keep going.Full definition → is now on by default: as a session nears automatic compaction, the summary is written in the background so the next prompt does not wait. An unused background summary still costs tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition →; `compaction.background: false` turns it off.
OpenCode again reads CLAUDE.md in any directory without an AGENTS.md, and falls back to ~/.claude/CLAUDE.md when the global AGENTS.md is missing. Setting OPENCODE_DISABLE_CLAUDE_CODE skips every CLAUDE.md.
Experimental schedules became automations. They fire on a time, interval, cron or pushed MCPThe Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.Full definition → event, then send a prompt or run a Code Mode program that decides whether to send one. Schedules with a condition or run limit, and finished ones, are dropped in the move.
Terms in this piece · Glossary
context compaction — Summarizing an agent's earlier conversation to free room in the context window so a long session can keep going.
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.