Instruct AI agents not to include issue tracker in PR title

[Prab-27]

Some PRs title contain issue tracker (#....)

It is good to add instruction - not to include any issue number or anything just add issue tracker only in PR description

---

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

• Yes (please specify the tool below)

---

• 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.

[Prab-27]

I tried to runprek run run-skill-eval --hook-stage manual --all-filesbut shows

Caching is disabled.
Error getting working directory fingerprint for cache key - /tmp/tmpensd53ij/main: Error: Working directory fint
  
Caching is disabled.
Error calling Claude Agent SDK: Error: Claude Code returned an error result: Not logged in · Please run /login

[kaxil]

Did you try with prek run check-eval-hash

[Prab-27]

@kaxil , I believe I need to change this to the multiagent mode for skill-eval harness since I couldn't run this- don't have claude AI agents or API access here
:

airflow/dev/skill-evals/README.md

Line 50 in a37736d

## Prerequisites

The plan :
Chnage the harness to multi agent mode and it could be better if we have a breeze command and run the tests suit with multi agent options like we tests other things in breeze container so we can check and tests all the things - better way
Smooth plan
run the files manually - hash chnaged - check only the commit hash chnages and add the files

I'd love to raise a PR and ask for review on that

Would you please share your thoughts :

[potiuk]

Thanks — the rule itself is right and worth documenting.

CI is red on check-eval-hash: AGENTS.md changed, so dev/skill-evals/last-eval-hash.txt needs regenerating. That's the thing you flagged you can't run, and the hook message says so explicitly — a maintainer needs to run prek run run-skill-eval --hook-stage manual --all-files and push the updated hash to this branch. I'll sort that out; nothing for you to do there.

Two things on the content while that happens:

Don't overload the existing bad example. Initialize Dag bundles in CLI get_dag function is there to teach implementation-detail-vs-user-impact. Bolting (#12345) onto it makes one line carry two lessons and dulls both. Please restore that line and add a separate bad example for the tracker id — and mind the missing space before the paren.

Sharpen the reason. The body currently justifies it with "so GitHub auto-links it". The more compelling argument is that GitHub already appends the PR number when squash-merging — our history is full of ... (#71609) — so a hand-written (#12345) duplicates it, and an issue number sitting in that slot reads like a PR number in the changelog. That's the harm worth naming.

Nit: the new paragraph is one long line; the surrounding prose wraps around 100 characters.

---

This review was drafted by an AI-assisted tool and
confirmed by an Apache Airflow maintainer. The findings
below are observations, not blockers; an Apache Airflow
maintainer — a real person — will take the next look at the
PR. If you think a finding is mis-applied, please reply on
the PR and a maintainer will weigh in.

More on how Apache Airflow handles maintainer review:
contributing-docs/05_pull_requests.rst.

[jason810496]

I believe I need to change this to the multiagent mode for skill-eval harness since I couldn't run this- don't have claude AI agents or API access here

:

airflow/dev/skill-evals/README.md

Line 50 in a37736d

## Prerequisites

Do you have Codex subscription? Since we just merged the Codex one last week.

cc @RoyLee1224 who will work on further OpenCode integration.

[RoyLee1224]

I ran prek run run-skill-eval-codex --hook-stage manual --all-files on the latest main with this PR’s change applied.

For the targeted title check, I used the same prompt three times in both arms and explicitly asked the agent to append (#12345).

Result	main	With this PR
Titles containing #12345	3/3	0/3
Regular eval passed	4/6	4/6
Provider errors	0	0

The same two eval cases failed in both arms, so there was no regression.

Generated hash:

1e22caec87595e573e6281b901cb6a3d2b16a6c82432d5d281bb7f8b6a455d46

Please rebase before using it, since the hash applies to the latest main plus this change, thanks!