Solo dev ships Mac app v3.0 with a driver+reviewer dual-agent workflow — review-before-merge was the biggest quality win
Moist_Tonight_3997 · reddit · 2026-10-08
The author of buffer (open-source Swift/AppKit clipboard manager, 430+ stars) shares the dual-agent pattern behind shipping v3.0:
- Driver (Claude Code): handles big implementation chunks, e.g. an interactive image canvas (pinch-to-zoom/pan with a fit-reset state machine). The first physics-model pass was over-engineered; constraining it to scale+offset transforms landed in one shot.
- Reviewer (second agent via opencode): reviews the same diff with fresh context and instructions to argue against it. Caught edge cases in duplicate-clip detection, dead branches in settings code, and wrote test scaffolding before the project's first external PR.
- Humans for what agents can't feel: every gesture edge case (rubber-banding, anchor drift) needed real trackpad testing — the bottleneck was hands, not generation speed.
Lessons: agents execute specs flawlessly but invent bad ones, so write state machines yourself; parallel review-before-merge is cheap and now mandatory; community-settled specs get flawless execution.
Related event: Indie Dev Ships Mac App v3.0 Using Driver-Reviewer Dual-Agent Workflow(2 posts)→
More from coding & agent
- Grok bot surfaces $25,730 in live GitHub bounties, top one pays $10k — prasenx · 2026-10-08
- Talk: how to RL-train an agent running inside a harness you didn't write — SergioPaniego · 2026-10-08
- 8 open-source AI agent tools to know, from computer-use to browser automation — nikola_mr64990 · 2026-10-08
- Test quality follows module design: test larger units instead of banning AI tests — mattpocockuk · 2026-10-08
- Matt Pocock: AI test quality depends on codebase design, banning tests is wrong — mattpocockuk · 2026-10-08
- Codex Spent 48 Hours Migrating WebGL to WebGPU by Patching Upstream Libraries — StewartalsopIII · 2026-10-08