content: audio_ducking now has SDK resources and a CLI command

[spencer-zqian]

Summary

The SDKs grew a dedicated audio-ducking resource (`client.audio_ducking` in Python 0.13.0 — sonilo-python#29, `client.audioDucking` in JS 0.14.0 — sonilo-js#51) and both CLIs grew `sonilo audio-ducking` (CLI 0.12.0), closing the one generation-tool gap this repo had to document around.

• audio-ducking/SKILL.md: the "no dedicated SDK method / `client.request()` escape hatch" section is replaced with ordinary Python / JS / CLI quick starts, matching the other skills' structure (install lines pin the first versions that carry the resource).
• tests/tool_surface.json: `audio-ducking` added to the CLI commands and `audio_ducking → audio-ducking` to the tool→command map; `audio_ducking` dropped from the `no_command` whitelist, so any future "no CLI command" claim about it fails CI instead of going stale in the other direction.
• README: the "not every tool has an SDK/CLI equivalent" caveat now covers only `play_audio`, which is local playback and MCP-only by nature.

⚠️ Merge after the SDK/CLI releases ship (sonilo-js#51 → npm `sonilo` 0.14 / `sonilo-cli` 0.13, sonilo-python#29 → PyPI `sonilo` 0.13 / `sonilo-cli` 0.12): the quick starts name those versions, and until they are on npm/PyPI the commands do not exist for users.

`python3 tests/validate.py` passes.

🤖 Generated with Claude Code

[sapient-app]

Lightsage docs evals

Waiting for the staging docs URL before running evals.

Lightsage will start the selected PR evals automatically when GitHub reports a successful docs deployment for this PR. This usually happens within 15 minutes.

Commit: fec46c7
Status: waiting for staging docs URL