Open-source coding agent cuts large-repo bugfix cost by up to 75% after learning the repo once
NeighborhoodOwn8510 · reddit · 2026-07-24
An open-source multi-agent SDLC harness claims it can beat a cold Claude Code run on large repositories by learning a repo once and reusing that knowledge.
- The author says AutoDev Studio was 7%–75% cheaper than a cold claude -p run across 6/6 localized tasks on repositories up to 82k LOC.
- Example cost: a cold agent run took $6.83 and 207 turns, while AutoDev Studio handled the same bug for about $1.70.
- The system persists repository knowledge with static analysis and a local embedding index, so future tasks use lookup instead of re-discovery.
- Its pipeline includes a PM agent for clarifying questions and tickets, a Dev agent on an isolated branch, QA test execution, a separate reviewer model family, bounded revision loops, and real GitHub PR creation.
- It also ships with a live Kanban board, token/cost tracking, offline default operation, FastAPI + SQLite, and MIT licensing.
- The author notes where it loses: tiny easy edits can be cheaper with a one-shot agent, and one cross-cutting bug got a narrower fix than the baseline.
Related event: Open-source AutoDev Studio Claims to Cut Bug Fixing Costs by Up to 75%(2 posts)→
More from coding & agent
- Krea 2 identity edit keeps the face but adds ghosting and blur — witcherknight · 2026-07-24
- Claude skills can fail even when the body is right, because routing depends on the description — socialwithaayan · 2026-07-24
- Claude skills should split prose, scripts, and tools instead of keeping everything as text — socialwithaayan · 2026-07-24
- Claude skills rot over time: a framework for auditing dead skills, collisions, and drift — socialwithaayan · 2026-07-24
- How to encode a writer’s voice into rules a fresh Claude session can reproduce — socialwithaayan · 2026-07-24
- How to chain Claude skills into a clean production line with strict handoff contracts — socialwithaayan · 2026-07-24