Add: lavs9/quantwave (agent skill for consistent research↔live TA/backtest)

[lavs9]

1. Repo URL · 仓库链接

https://github.com/lavs9/quantwave

2. Pillar + sub-category · 放在哪个分类

Skills → Strategy coding & backtesting

3. Proposed entry bullet · 建议的条目

- [lavs9/quantwave](https://github.com/lavs9/quantwave) - Polars-native TA + backtest library with batch==streaming parity and an agent skill so coding agents keep research and live strategy code consistent.

Chinese (approx):

- [lavs9/quantwave](https://github.com/lavs9/quantwave) - 面向 agent 的 Polars 原生 TA/回测库：batch 与 streaming 位一致，并附带 agent skill，降低 research→live 数值漂移。

4. Metadata at proposal time · 提交时的基本信息

• GitHub stars: ~6 (below the default 100-star floor — opening as issue per CONTRIBUTING for early projects)
• Last commit or activity date: 2026-08 (active)

5. Quality-bar self-evaluation · 条目自查

• Open source (MIT)
• Demonstrably LLM-driven: ships a Claude/agent skill (SKILL.md + pitfalls + usage linter) so agents write against real TA/backtest conventions; not a classical-only lib without agent surface
• Active (recent commits)
• Clear documentation (MkDocs site + agent skill guide)
• Distinct contribution: research↔live numerical consistency + Polars expression plugins + agent skill (not another multi-agent trader)
• 100+ stars: not yet — requesting maintainers' judgment / watchlist rather than force-merge

6. Why it fits this list (not only awesome-quant)

Classic TA libs belong on awesome-quant (we are also submitting there). The agent skill is the reason for this list: coding agents generate strategy code that silently forks between notebook, backtest, and live. QuantWave's skill + parity design is aimed at that failure mode.

Docs: https://lavs9.github.io/quantwave/guides/agent-skill/

Happy to wait until stars clear the floor, or to rephrase if maintainers prefer a different section.