← ClaudeAtlas

follow-the-threadlisted

When a red CI test, a failing check, or a single bug crosses your path, follow it instead of dismissing it — one thread reliably sources many issues worth working. A failing test maps to an existing issue whose root cause is often an existing UPSTREAM bug (downstream symptom → upstream cause); fixing that bug means reading the surrounding subsystem, where adjacent open issues surface; searching near the one you're fixing (same area / same reporter) seeds the next. A convenient "it also fixes X" is a hypothesis, not a finding — verify the mechanism. This is a SOURCING method: it finds where the help actually is, before any PR. Use when you hit a red CI on your own PR, a flake, or one bug and want to find the real work around it. Trigger terms: red CI, flaky test, follow the thread, downstream symptom, upstream root cause, adjacent issue, sibling bug, "not my bug", where's the real work.
serhiy-bzhezytskyy/contrib-receipts · ★ 1 · AI & Automation · score 77
Install: claude install-skill serhiy-bzhezytskyy/contrib-receipts
# Follow the thread ## Purpose The most productive way to find work worth doing is not triage or a good-first-issue scan — it's to **follow a thread you already touched**. One red CI test, followed honestly instead of waved off as "not my bug," root-causes to an upstream issue; fixing that means reading the subsystem around it, where adjacent open issues are already filed; searching near that one surfaces a third. The signal is in the code you're already standing in. Etiquette skills stop you from being noise; this one finds where the signal is. The discipline is one question at every hop: **"why does THIS happen?"** instead of "not mine." Each honest answer either ends the thread or hands you the next issue. ## When to use - A red CI or a flake appears on your own PR, or on a repo you're already in. - You've root-caused one bug and want to know whether it's isolated or a nest. - You're standing in an unfamiliar subsystem and want to read outward, not just fix the one line you came for. ## When NOT to use You already have a scoped, agreed task to land — don't let thread-following become scope creep that stalls the PR in front of you; source the next thread after this one ships. ## The practice (checklist) - [ ] **Red CI on your own PR → follow it, don't dismiss it.** A flake on your PR is a seam. Triage it honestly first: is it actually flaky, or deterministic? - [ ] **Map the failing test to its existing issue, then to the UPSTREAM cause.** A downstrea