Skip worktrees: clone repos per work stream and ship layered mega PRs
idanbeck · x · 2026-09-13
Responding to Eric Friedman's lost-worktree incident with Codex, idanbeck explains why he's "super negative" on worktrees: agents working in them always end in lost data or forensic reconstruction mess. His alternative pattern — a plain repo clone for each work stream organized in a monorepo hierarchy, merged via mega PRs with layered commits. First-hand experience in the worktree-vs-clones debate for parallel agent workflows.
Related event: Codex Overnight Session Wrecks Worktree; Devs Suggest Repo Clones Instead(3 posts)→
More from coding & agent
- Testing agents with explicit rejection records: a cheaper path to verified troubleshooting answers — krishnan · 2026-10-31
- krishnan proposes logging ruled-out troubleshooting hypotheses for future reopening — krishnan · 2026-10-31
- krishnan on autonomous lab agents: criticism changed the next search, not just verdicts — krishnan · 2026-10-31
- Multi-agent AI lab autonomously researches quantum spins in diamond, revises drifting experiments — krishnan · 2026-10-31
- Next 9 proposals explored 7 new directions: rejection became information — krishnan · 2026-10-31
- PI agent rejected 75 proposals in 30 hours; manager banned recurring bad ideas — krishnan · 2026-10-31