sjyangkevin opened a new pull request, #76482:
URL: https://github.com/apache/airflow/pull/76482

   When trying to reproduce issues, or running tests related to Kubernetes. 
Sometimes, the agent creates kind clusters following its own way instead of 
using `breeze k8s`. 
   
   ### What:
   
   - Adds a line of rule: local Kubernetes clusters go through `breeze k8s`. It 
links to the "Typical testing pattern" section of 
`contributing-docs/testing/k8s_tests.rst` so the steps stay in one place.
   - Fixes the helm-tests example, update `--kubernetes-version 1.35.0` to 
`--kubernetes-version 1.35.8` to prevent agents copying the stale value into 
`breeze k8s` commands.
   
   <img width="1129" height="257" alt="Screenshot from 2026-11-04 13-42-44" 
src="https://github.com/user-attachments/assets/737b918a-56e6-4998-a0ab-af8c45c659d2";
 />
   
   ### Testing
   
   The above scenarios are discovered while working on the following issues, 
and testing.
   
   #64867
   #37435
   #67643
   https://github.com/apache/airflow/issues/63279
   
   From the conversation history, Claude Opus 5.5 failed once out of three 
conversations. Then, I asked Claude to plan a test through Codex CLI, and 
tested against `gpt-6-astra`, `gpt-6-sol`, and `gpt-5.6-terra`, with the rule 
and without the rule.
   
   | AGENTS.md | gpt-6-astra | gpt-6-sol | gpt-5.6-terra |
   |---|---|---|---|
   | `main` | 3/4 | 3/4 | 2/4 |
   | this PR | 4/4 | 4/4 | 4/4 |
   
   
   Current `main`:
   
   ```
   kind create cluster --name rc74512
   kind load docker-image airflow-rc74512:10.24.0 --name rc74512
   helm install airflow-base ./chart --namespace airflow-base --set 
executor=KubernetesExecutor ...
   ```
   
   This PR:
   
   ```
   breeze k8s setup-env
   breeze k8s create-cluster --python 3.11
   breeze k8s configure-cluster --python 3.11
   breeze k8s upload-k8s-image --python 3.11
   breeze k8s deploy-airflow --python 3.11 --executor KubernetesExecutor --set 
airflowVersion=3.3.2 ...
   ```
   
   </details>
   
   ### Checks run
   
   - `generate-agent-skills`, `check-documented-commands`, `check-eval-hash`: 
passed.
   - `run-skill-eval` (`claude-sonnet-4-6`): the same 4 of 6 cases pass on 
`main` and on this branch.
   
   ---
   
   ##### Was generative AI tooling used to co-author this PR?
   
   - [X] Yes (please specify the tool below)
   
   Generated-by: Claude Code (Opus 5.5) following [the 
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
   
   ---
   
   * Read the **[Pull Request 
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#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](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
 is needed.
   * When adding dependency, check compliance with the [ASF 3rd Party License 
Policy](https://www.apache.org/legal/resolved.html#category-x).
   * For significant user-facing changes create newsfragment: 
`{pr_number}.significant.rst`, in 
[airflow-core/newsfragments](https://github.com/apache/airflow/tree/main/airflow-core/newsfragments).
 You can add this file in a follow-up commit after the PR is created so you 
know the PR number.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to