bailian-tts

Generate speech audio with 阿里云百炼 TTS via the `bailian-cli` npm package. Use when users ask to convert text to voice, choose voices/languages, batch-generate...

Security Scan

The skill appears to do what it says (call a Bailian TTS CLI), but it omits declaring the required API credential and relies on an unverified npm package that the quick-test will install globally — these mismatches are worth investigating before installing.

Purpose & Capability

The name/description match the instructions: it drives a bailian-cli to produce TTS. However, the skill fails to declare the main runtime credential (BAILIAN_API_KEY is checked/required in SKILL.md and the test script but requires.env lists none), which is an incoherence between claimed metadata and actual needs.

Instruction Scope

Runtime instructions and the included quick-test.sh are limited to checking/installing the CLI, verifying an API key, listing voices, and invoking bailian tts. They do not instruct reading unrelated system files or exfiltrating data. They do offer a base64 output option and default an output folder under ~/.openclaw/media/audio, which is reasonable for TTS workflows.

Install Mechanism

There is no formal install spec, but the provided script and instructions run npm i -g @hackerpl/bailian-cli. That is a global install of an npm package with no homepage/repository provided in the skill metadata and no verification of provenance. Installing a third-party global CLI is moderate risk and should be verified (package ownership, source repo, audit of package contents).

Credentials

The skill actually requires BAILIAN_API_KEY (and optionally BAILIAN_REGION) for operation, but the skill metadata lists no required env vars or primary credential. Asking for an API key is proportionate for a TTS integration, but the omission in the declared requirements is an inconsistency that reduces transparency. The script also assumes standard HOME/OUT_DIR locations, which is normal.

Persistence & Privilege

The skill does not request permanent presence (always:false), does not modify other skills or system-wide agent settings, and does not appear to persist credentials itself. The only system change the instructions suggest is installing a global npm CLI and creating an output directory.

What to consider before installing

This skill looks functionally coherent for TTS but has two practical concerns you should resolve before installing: (1) it requires a BAILIAN_API_KEY at runtime (the SKILL.md and quick-test.sh check for it), but the skill metadata doesn't declare that credential — expect to provide that secret for use; (2) the CLI it invokes is @hackerpl/bailian-cli (no homepage/repo in metadata). Verify the npm package is legitimate: check the package page, repository, maintainer identity, download counts, and review its code (or prefer an officially maintained Alibaba/百炼 client). Note the quick-test script will auto-install the CLI globally if missing — avoid running it on production hosts or run it in an isolated environment/container. Finally, avoid sending highly sensitive text to the TTS service unless you confirm the API key scope and the service's privacy terms.

Use this skill to produce speech audio from text with @hackerpl/bailian-cli.

Quick Workflow

1. Check CLI availability.
2. Check BAILIAN_API_KEY.
3. If key is missing, guide user to prepare one in 阿里云百炼.
4. Run bailian tts with requested voice/language/style.
5. Return output path (or base64 when requested). Default audio output directory: ~/.openclaw/media/audio.

1) Check and install CLI

Run:

npm ls -g --depth=0 @hackerpl/bailian-cli

If missing, install:

npm i -g @hackerpl/bailian-cli

2) Check API key

Run:

[ -n "$BAILIAN_API_KEY" ] && echo SET || echo MISSING

If missing, ask user to create and configure key from the 阿里云百炼官方入口（控制台/平台）：

• https://bailian.console.aliyun.com/

Then set it:

export BAILIAN_API_KEY="sk-xxxx"
# optional
export BAILIAN_REGION="beijing"   # or singapore

3) Generate speech (bailian-cli)

Basic (default to OpenClaw media folder):

mkdir -p ~/.openclaw/media/audio
bailian tts -t "你好，欢迎使用百炼 TTS 服务" -o url -d ~/.openclaw/media/audio

Specify voice/language:

mkdir -p ~/.openclaw/media/audio
bailian tts -t "Hello world" -v "Ethan" -l "English" -o url -d ~/.openclaw/media/audio

With style instruction:

mkdir -p ~/.openclaw/media/audio
bailian tts -t "春眠不觉晓" -v "Serena" -i "用温柔缓慢的语调朗读" -o url -d ~/.openclaw/media/audio

Return base64 instead of file:

bailian tts -t "测试" -o data

Custom output directory:

bailian tts -t "你好世界" -d "./audio"

4) Voices and languages

• For live lookup:

bailian tts --list-voices

• Full built-in table: read references/voices-and-languages.md.

5) Operational notes

• Keep text length within CLI limit (600 chars per request).
• Default output directory: ~/.openclaw/media/audio (create with mkdir -p ~/.openclaw/media/audio).
• Prefer -o url for file delivery, -o data for programmatic pipelines.
• Respect user privacy: do not upload sensitive text externally without explicit confirmation.
• If generation fails, check in order: API key → region → network → voice/language spelling.
• When producing many clips, keep a consistent voice and instruction style for tonal continuity.