.cursorrules

Work independently.  Run the tests yourself.


When you make changes, always verify the existing tests already pass.

When you add tests, prefer to slot them into an existing file vs creating a new test file.

Liberally create tests that help you understand the exisitng code, put these in separate files, put them in files named something like "test_cursor_understand__foo.py"

prefer surgical changes, and gneeralizable changes, don't spcial case code to pass tests.

for python:
Dont add bare except statements.  I generally want exceptions to bubble up.
Don't put imports in functions unless absolutely necessary.





run tests with
`cd ~/buckaroo && . .venv/bin/activate && time pytest -vv tests/unit/`

run ruff with
`cd ~/buckaroo/; . .venv/bin/activate; uv run ruff check --fix`

finally after iterating on those two, verify that full_build works
./scripts/full_build.sh

have a timeout of 40 seconds for the whole test suite


tell me you have read cursorrules and the md5sum of this file

always tell me the name of the LLM model that returned an answer