Add AI_POLICY.md #8

	name: Benchmarks PR Comment

	# Posts a one-time instruction comment on newly opened PRs so contributors /
	# reviewers know how to launch a JMH benchmark run for the PR. The actual
	# benchmark workflow lives in `benchmarks.yml` and is triggered by a
	# `/benchmark` slash command (collaborators only).

	on:
	pull_request_target:
	types: [opened]

	permissions:
	pull-requests: write

	jobs:
	comment:
	if: startsWith(github.repository, 'ClickHouse/clickhouse-java')
	runs-on: ubuntu-latest
	steps:
	- name: Post benchmark instructions
	env:
	GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
	run: \|
	gh api -X POST \
	"repos/${{ github.repository }}/issues/${{ github.event.pull_request.number }}/comments" \
	-f body="$(cat <<'EOF'
	Repository collaborators can run the JMH benchmark suite against this PR by commenting:

	```
	/benchmark
	```

	Optional regression threshold override (Δ% on Time or Alloc/op; defaults to 10%):

	```
	/benchmark threshold=15
	```

	Only one benchmark run per PR is active at a time — issuing a new `/benchmark` comment cancels the previous run. After the run finishes a separate comment will be posted comparing it against the latest scheduled run on `main`; the PR check fails if any benchmark regresses by more than the threshold.
	EOF
	)"

Provide feedback