Problem: Kattis and USACO had zero offline test coverage — no fixtures,
no conftest routers, and no entries in the test matrix. Precision, error
cases, and the Kattis contest-vs-slug fallback were also untested.
Solution: Add HTML fixtures for both platforms covering metadata, tests,
and contest list modes. Wire up conftest routers that patch
`httpx.AsyncClient.get` using the same pattern as CSES/CodeChef. Extend
the test matrix to include Kattis and USACO (18 parametrized cases, up
from 12). Add a dedicated test for the Kattis contest-path (verifies
`contest_url`/`standings_url` are set). Add parametrized metadata error
tests for CSES, USACO, and Kattis. Assert `precision` field type in all
tests-mode payloads; `usaco/problem_1471.html` includes an absolute-error
hint to exercise `extract_precision`.
## Problem
`codeforces.py` used `curl_cffi` to bypass Cloudflare when fetching
contest problem HTML, making it unavailable in the nix python env and
requiring an extra dependency across `pyproject.toml` and `flake.nix`.
## Solution
Rewrite `_fetch_problems_html` to use scrapling `StealthySession` with
`solve_cloudflare=True`, matching the existing CF submit pattern. Extend
`needs_browser` in `scraper.lua` to route CF `metadata` and `tests`
through the FHS env on NixOS. Remove `curl-cffi` from `pyproject.toml`,
`flake.nix`, and test mocks.
