drive: cloud run 1ffd2aee

[kjgbot]

Automated drive work from cloud run 1ffd2aee-035f-4410-8590-d9147205e132.

The sandbox cannot open PRs (no remote, no GitHub token), so this was delivered
from a host that can. Verification and adversarial review ran in-run — see
ops/reviews/ in the diff. A human merges.

[coderabbitai]

Warning

Review limit reached

Next included review available in 16 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Free

Run ID: 3cc07a26-ccce-4bdb-b3ab-91c2f5a86ba8

📥 Commits

Reviewing files that changed from the base of the PR and between 5cc0b2a and 7c6fe6f.

📒 Files selected for processing (3)

• ops/NEXT.md
• workflows/drive-cloud.yaml
• workflows/drive.yaml

---

Note

🎁 Summarized by CodeRabbit Free

Your organization is on the Free plan. CodeRabbit will generate a high-level summary and a walkthrough for each pull request. For a comprehensive line-by-line review, please upgrade your subscription to CodeRabbit Essentials by visiting https://app.coderabbit.ai/settings/billing.

_{Comment @coderabbitai help to get the list of available commands.}

[kjgbot]

Checked the risky part of this diff; it is correct

The change that deserved scrutiny is the widened tree-slim:

-rm -rf sdk/node_modules 2>/dev/null || true
+rm -rf sdk/node_modules sdk/dist surface/dist 2>/dev/null || true

Adding sdk/dist to a delete list is the kind of thing that has already cost this repo real time — an absent sdk/dist produced 18 test failures that were blamed on product code earlier this cycle, because the live-kernel suite requires the built CLI at sdk/dist/cli.js. So the comment's claim, "nothing downstream of here reads them," is load-bearing and I checked it rather than taking it.

It holds. After the slim line, exactly one thing runs in drive.yaml:

node ops/open-pr.mjs

and that script imports only node:child_process and node:fs, with no reference to sdk/dist, surface/dist, or node_modules. The suites and NEXT validation are genuinely finished by that point, and the trees are gitignored so the commit/handoff path never touches them.

Verdict: safe, and worth landing. It attacks a real, well-documented failure — the 413 flush that silently lost five runs' work — and it does it in the one place where deleting build output cannot affect a verdict, since the cleanup is guarded with || true and placed after the pass/fail decision is computed.

One note for whoever measures the result: the existing comment says sdk/node_modules was 1226 files and 52 MB. It would be worth recording what the census prints after this change, because the same comment block documents three prior wrong hypotheses about the 413 — and this is the fourth intervention against it. If the count does not move much, that is a signal in itself.

Unrelated: ops/NEXT.md is now a four-way collision

#189, #194, #199 and this PR all rewrite that file. Whichever lands last conflicts with the rest, and a merger resolving it by taking one side wholesale discards the others silently. That needs consolidating before any of them merges — it is now the most likely way for work to be lost in this repo.