Fix documented Breeze selective checks command

[viiccwen]

closes #69858

Correct the contributor documentation to use breeze ci selective-check, which is the command registered by Breeze and executed by the GitHub Actions workflows.

This updates the source AGENT-SKILL command block, regenerates the Commands section in AGENTS.md, fixes the local reproduction example in the selective checks troubleshooting guide, and adjusts the generator test to verify the valid command.

---

Was generative AI tooling used to co-author this PR?

• Yes - GPT-5.6 Sol

---

• Read the Pull Request Guidelines for more information. Note: commit author/co-author name and email in commits become permanently public when merged.
• For fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
• When adding dependency, check compliance with the ASF 3rd Party License Policy.
• For significant user-facing changes create newsfragment: {pr_number}.significant.rst, in airflow-core/newsfragments. You can add this file in a follow-up commit after the PR is created so you know the PR number.

[viiccwen]

Sorry I don't have Claude 😭
Does anyone can help me to update last-eval-hash.txt?

[Andrushika]

Hi! I have regenerated the hash for you. You can simply paste it to dev/skill-evals/last-eval-hash.txt, then CI should pass.
Hash: cdf346752bd43c958355d98e63453bc4a00dd82727dde4243b1c01a2e2afb5fe

[viiccwen]

Hi! I have regenerated the hash for you. You can simply paste it to dev/skill-evals/last-eval-hash.txt, then CI should pass. Hash: cdf346752bd43c958355d98e63453bc4a00dd82727dde4243b1c01a2e2afb5fe

Thx for your kindness. 🫡

[nailo2c]

Thanks, really helpful 😄

[RoyLee1224]

Thanks for the fix, and sorry the eval-hash gate (from #69308) added friction here. I'm experimenting with additional agent runtimes for the eval (Codex and OpenCode SDKs)and will propose that in a follow-up PR, so the gate won't assume a Claude subscription. 🙌

[jason810496]

Thanks for the fix, and sorry the eval-hash gate (from #69308) added friction here. I'm experimenting with additional agent runtimes for the eval (Codex and OpenCode SDKs)and will propose that in a follow-up PR, so the gate won't assume a Claude subscription. 🙌

Nice!

[jason810496]

Thanks for the catch.

Additionally, this is exactly what #64012 would like to prevent from happening -- we should introduce an automated mechanism to prevent further drift instead of relying on manual catch.

[github-actions]

Backport failed to create: v3-3-test. View the failure log Run details

Note: As of Merging PRs targeted for Airflow 3.X
the committer who merges the PR is responsible for backporting the PRs that are bug fixes (generally speaking) to the maintenance branches.

In matter of doubt please ask in #release-management Slack channel.

Status	Branch	Result
❌	v3-3-test

You can attempt to backport this manually by running:

cherry_picker d9dcc11 v3-3-test

This should apply the commit to the v3-3-test branch and leave the commit in conflict state marking
the files that need manual conflict resolution.

After you have resolved the conflicts, you can continue the backport process by running:

cherry_picker --continue

If you don't have cherry-picker installed, see the installation guide.

[RoyLee1224]

I was curious what a wrong documented command like this actually costs an agent in tokens, so I ran this fix through the skill-eval harness (#69308) as a before/after A/B — same task ("determine which CI test types run for HEAD"), same tools, only AGENTS.md differs, n=3 per arm:

Every "before" run trusted the doc, failed, and had to rediscover breeze ci selective-check via --help — roughly 2× tokens per affected task, re-paid on every fresh session. And that's the best-case branch: when the agent doesn't recover, the cost is way more than 2×.

[RoyLee1224]

@jason810496 Agreed! #68204 only keeps docs and AGENTS.md in sync, the bug was in the docs source, so the sync faithfully copied it over. Nothing checks that the commands are actually real.

Working on that now: validate documented commands against the real Breeze CLI, both when docs are generated and when a command gets renamed.