fix: improve research bot implementation quality #2

	name: CI

	on:
	push:
	branches: [ main, master, "**" ]
	pull_request:

	jobs:
	lint-and-test:
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v4
	- uses: actions/setup-python@v5
	with:
	python-version: '3.11'
	- name: Install dev deps
	run: \|
	python -m pip install --upgrade pip
	pip install pytest ruff
	- name: Ruff
	run: \|
	ruff --version
	ruff check scripts tests
	- name: Pytest
	run: \|
	pytest -q tests

	bot-idempotency:
	runs-on: ubuntu-latest
	if: ${{ github.event_name == 'pull_request' && startsWith(github.head_ref, 'research-bot/') }}
	steps:
	- uses: actions/checkout@v4
	with:
	fetch-depth: 0
	- uses: actions/setup-python@v5
	with:
	python-version: '3.11'
	- name: Run bot and verify no diff
	run: \|
	python scripts/research_bot.py \|\| true
	git status --porcelain
	# Ensure the bot is idempotent for its own PR branch
	if ! git diff --quiet; then
	echo "Bot produced changes on its own PR branch. Please re-run the bot locally and commit the updates." >&2
	exit 1
	fi

Provide feedback