Four open-source AI agents killed mid-action all repeat side effects on retry
merlinofthewater · reddit · 2026-10-09
The author crashed four published agent codebases (Cua, Skyvern, Corsair, Inconvo) at pinned commits mid-action and retried like a supervisor would: Cua double-charges on checkout, Skyvern strands tasks in 'running' forever, Corsair corrupts its rotating OAuth token, and Inconvo pays 60% more model calls per interrupted answer. Shared root cause: none records which actions already completed. A 25,930-episode arXiv study found agents self-report success in 90% of episodes where they duplicated an effect, so self-reported traces can't be trusted. Harnesses and raw ledgers are open-sourced, running against local fakes with no API key needed.
Related event: Four Open-Source Agents All Repeat Completed Actions After Interruption(2 posts)→
More from coding & agent
- Self-organizing agent teams hit 66.7% on math benchmarks vs 48.8% for their strongest member — zainhas · 2026-10-12
- Practitioner's verdict on autoresearch: great at speeding experiments, not at frontier runs — iaindunning · 2026-10-12
- levelsio cancels nearly all SaaS subscriptions: money now goes only to AI inference, content, data centers, power and taxes — karlwaldman · 2026-10-12
- Grok Bot is winning the agentic personal assistant race, says dev — Arindam_1729 · 2026-10-12
- Does Codex refuse or cut back work when its own budget estimate runs high? — r618NecessaryStation · 2026-10-12
- Excalidraw for slides is the GOAT: MCP integration plus animations, no AI slop — HamelHusain · 2026-10-12