/target
**/*.rs.bk
Cargo.lock
# python build / artifacts
*.so
*.pyd
__pycache__/
.venv/
*.npy
/dist/
*.whl
.coverage
.coverage.*
.mypy_cache/
.ruff_cache/
.pytest_cache/
# docs site build
/site/
# local planning / research notes / scratch measurements (not shipped)
/plans/
/local/
# agent rules, memory pointers and session state -- local only, never published
/CLAUDE.md
/.claude/
/AGENTS.md
/.cursorrules
# editor
.idea/
.vscode/
# JOSS paper build artifact (paper.md / paper.bib are tracked)
/paper/paper.pdf

# Per-seed benchmark runs: intermediate artifacts of `bench/median_of_seeds.py`. The published
# medians (results_*.csv) and their min/median/max (results_*_spread.csv) are committed; regenerate
# these with `comprehensive.py --seed S --tag _sS` if you want to audit the reduction.
bench/results_*_s[0-9].csv

# cargo-mutants run artefacts (hundreds of MB; the accepted survivors live in
# mutants-baseline.txt, which is tracked)
mutants.out/
mutants.out.old/
