Adopt uv for Python dependency management — replace manual pip/venv setup with one `uv sync`

[AsharibAli]

Problem

Setting up the toolkit's Python side currently requires several manual steps that vary by platform:

• python3 -m pip install -r tools/requirements.txt for the base dependencies
• Manually creating/activating a virtualenv (docs show python -m venv .venv + source .venv/bin/activate)
• pip install modal && python3 -m modal setup as a separate step for the Modal CLI
• pip install openai-whisper hidden in a commented-out line of requirements.txt for karaoke captions
• --break-system-packages workarounds on Debian/Ubuntu (PEP 668) in the OpenClaw skill

There's also no lockfile, so installs aren't reproducible, and users on a machine without the right Python version have to install one themselves first.

Proposal

Adopt uv as the Python project manager:

• A root pyproject.toml + uv.lock so the entire setup becomes one command: uv sync
• Optional heavy/niche deps become extras: uv sync --extra whisper, uv sync --extra modal
• All tool invocations documented as uv run tools/<tool>.py — no venv activation, auto-syncs the env
• uv installs a compatible Python automatically (.python-version pins 3.12)
• 10–100x faster installs than pip, fully reproducible via the lockfile
• tools/requirements.txt kept as a pip fallback for anyone not using uv

Out of scope: docker/** images (containers manage their own deps) and the changelog history.

I have this working end-to-end and will open a PR.

[AsharibAli]

I have implemented the proposed changes in the following pull request: #33