Open every PR with which checks are safe to merge without

[Peeja]

⏭️ Safe to merge without waiting for CI — all of it

Refreshed on every push. Checks are not required yet, so this is advice, not a gate.

One file changed: AGENTS.md. Nothing in the repository reads it.

$ git diff --name-only origin/main HEAD
AGENTS.md
$ grep -rn "AGENTS.md" --include='*.yml' --include='*.sh' --include='*.go' --include=Makefile .
(no output)

No workflow, guard script, Go package or Makefile target mentions it. ci, images, e2e and itest will all pass, and none of them could have done otherwise.

What this costs if it's wrong: the checks still run; we just don't wait. If an ignored check goes red after merge, main is red. That is the trade, and it is why the reasoning is re-derived per push rather than encoded as a paths: list that goes stale.

---

Your idea, written down so it survives this session. One file, based on main, stacked on nothing.

The problem it addresses

Nothing is path-filtered, on purpose, and itest alone is ~23 minutes. So a Markdown-only pull request — which most of the recent ones have been — costs a full run before anyone can merge it. #8 changed one Markdown file and still cost itest 26m27s.

Why manual beats the automatic version, for now

MONOREPO_TODO.md records the two faults real path filtering has here, and a per-PR block has neither:

	paths: filter	a block written per push
goes stale when a module gains a dependency	yes, silently	cannot — it is rewritten each time
a filtered job reports skipped, never satisfying a required check	yes	n/a, nothing is skipped

And it can say things no path pattern can express. The block on #21 right now says "itest already passed on f4c5c21f and git diff f4c5c21f HEAD -- .github/ is empty" — that is not a path, it is a fact about two shas.

The part that matters: derive, don't assert

The section names how, per change shape — Markdown only, one workflow's own YAML, Go code (go list -deps), already-green-at-an-earlier-head — because the closure here is not obvious, which is the entire reason CI is unfiltered. ingot → indexing-service and delegator → forgectl were both invisible in go.mod and only appeared under go list -deps (rule 3). A hand-waved "this can't affect anything" is how that gets missed.

It also requires the block to state what it costs when wrong. The checks still run; skipping the wait means an ignored check that goes red leaves main red. The reviewer is the one taking that trade and should be able to see it.

The blocks are kept, not discarded

When real filtering is designed, they are the worked examples of what it has to be able to express — and the ones that turned out wrong are worth more than the ones that didn't. That is the cheapest way to get evidence for a design nobody can currently specify.

Listed in the scaffolding section among the things a finished monorepo should drop, since it is explicitly an interim.

Worth your eye

This is a process proposal, like #20 — it changes how pull requests are written, not what the code does. Two things I'd look at: whether "derive, don't assert" is stated strongly enough to actually bind, and whether the four change-shapes cover the cases you expect (a Dockerfile-only change and a go.mod bump are both absent, deliberately — I did not want to guess at shapes I have not had to reason about yet).

🤖 Generated with Claude Code

https://claude.ai/code/session_01CGAGAib517Ae1kg8SCdcEt

---

Generated by Claude Code