Dev shares multi-agent coding stack: routing tasks to Opus, Fable and Astra by code size
sh_reya · x · 2026-10-26
shreya lays out her current intelligence stack for software development: humans create issues and feature ideas; human + Opus 5.5 iterate on RFCs extensively; tasks estimated at over 4k lines of code go to Fable, otherwise Opus 5.5 implements; Astra reviews PRs for bugs and re-invokes the implementation agent as needed; the human reviews code and proposes simplifications (goal: minimize lines added, maximize lines deleted) which Astra carries out; human + Sol handle documentation. The core pattern: humans stay on requirements, judgment and simplification, while specialized agents handle implementation, review and docs.
More from coding & agent
- Resin: a runtime that lets agents write forward passes to build standalone LLM inference engines — andersonbcdefg · 2026-10-26
- Developers share the hard rules they added to agents after a single bad experience — Luvena21 · 2026-10-26
- unified-brain: OSS SQLITE memory layer claims frontier-class coding at 1/18th cost — StatusFoundation5472 · 2026-10-26
- App Store Connect CLI 5.17.0 ships one-call screenshot reordering, 8k stars on GitHub — rudrank · 2026-10-26
- Dev shows GPT-6 Astra Ultra coding for 41+ hours straight on a Dreamcast game — AIandDesign · 2026-10-26
- Bitpanda crypto exchange gets an MCP server for trades, wallets and transactions — modelcontextprotocol · 2026-10-26