Alicloud Ai Audio Asr

Transcribe non-realtime speech with Alibaba Cloud Model Studio Qwen ASR models (`qwen3-asr-flash`, `qwen-audio-asr`, `qwen3-asr-flash-filetrans`). Use when c...

Security Scan

The skill mostly does what it says (Alibaba Cloud Qwen ASR transcription) but its declared metadata omits required credentials and the runtime script reads local .env and credentials files (potentially exposing unrelated secrets), so the package's claims and actual behavior are inconsistent.

Purpose & Capability

The skill's description and SKILL.md describe using Alibaba Cloud DashScope/Qwen ASR and require an API key (DASHSCOPE_API_KEY). However, the skill metadata declares no required environment variables or primary credential. This mismatch (no declared credential but runtime requiring one) is incoherent: a transcription skill legitimately needs an API key, so the metadata should list it.

Instruction Scope

The SKILL.md and bundled script instruct the agent to read environment variables, ~/.alibabacloud/credentials, and local .env files (current working directory and repository root). While reading an ASR API key is expected, automatically loading arbitrary .env files and repo-level .env can pick up unrelated secrets or configuration and incorporate them into the environment used for requests. The script also will base64-encode and upload local audio (data URI), which is expected for local-file transcription but is a potential data exfiltration pathway if users are unaware.

Install Mechanism

No install spec is provided (instruction-only with a helper Python script). The SKILL.md suggests creating a virtualenv but does not download external code or archives. The script uses only Python stdlib, so there is no high-risk install step.

Credentials

The runtime expects DASHSCOPE_API_KEY (and supports ALIBABA_CLOUD_PROFILE/ALICLOUD_PROFILE) but the registry metadata lists no required env vars or primary credential. The script will also load .env files from cwd and repository root and read ~/.alibabacloud/credentials to populate DASHSCOPE_API_KEY. Requiring the ASR API key is proportional to the stated purpose, but silently loading additional .env files and credentials is broader than necessary and not declared.

Persistence & Privilege

The skill does not request always:true and does not attempt to modify other skills or system-wide agent settings. It writes outputs to specified output paths under output/alicloud-ai-audio-asr/, which is consistent with its stated behavior.

What to consider before installing

This skill appears to be a legitimate Alibaba Cloud Qwen ASR helper, but its metadata omits the fact that it needs DASHSCOPE_API_KEY and that the bundled script will read .env files (cwd and repo root) and ~/.alibabacloud/credentials. Before installing: - Expect to provide DASHSCOPE_API_KEY (or add dashscope_api_key to ~/.alibabacloud/credentials). Prefer setting the env var at runtime rather than leaving keys in repo .env files. - Be aware the helper will read .env files and the repo root .env (if a .git directory exists). Remove or move any unrelated secrets from those .env files to avoid accidental use. - The script will base64-encode and upload local audio as data URIs; do not transcribe sensitive audio unless you trust the remote service and network. - Confirm the endpoints (dashscope.aliyuncs.com) and API behavior match your expectations and that you are comfortable sending audio and transcripts to that service. - If you require strict least privilege or want the metadata to be accurate, ask the publisher to declare DASHSCOPE_API_KEY (primary credential) in the skill manifest and to stop implicitly loading repository .env files or to make that behavior opt-in.

Category: provider

Model Studio Qwen ASR (Non-Realtime)

Validation

mkdir -p output/alicloud-ai-audio-asr
python -m py_compile skills/ai/audio/alicloud-ai-audio-asr/scripts/transcribe_audio.py && echo "py_compile_ok" > output/alicloud-ai-audio-asr/validate.txt

Pass criteria: command exits 0 and output/alicloud-ai-audio-asr/validate.txt is generated.

Output And Evidence

• Store transcripts and API responses under output/alicloud-ai-audio-asr/.
• Keep one command log or sample response per run.

Use Qwen ASR for recorded audio transcription (non-realtime), including short audio sync calls and long audio async jobs.

Critical model names

Use one of these exact model strings:

• qwen3-asr-flash
• qwen-audio-asr
• qwen3-asr-flash-filetrans

Selection guidance:

• Use qwen3-asr-flash or qwen-audio-asr for short/normal recordings (sync).
• Use qwen3-asr-flash-filetrans for long-file transcription (async task workflow).

Prerequisites

• Install SDK dependencies (script uses Python stdlib only):

python3 -m venv .venv
. .venv/bin/activate

• Set DASHSCOPE_API_KEY in environment, or add dashscope_api_key to ~/.alibabacloud/credentials.

Normalized interface (asr.transcribe)

Request

• audio (string, required): public URL or local file path.
• model (string, optional): default qwen3-asr-flash.
• language_hints (array<string>, optional): e.g. zh, en.
• sample_rate (number, optional)
• vocabulary_id (string, optional)
• disfluency_removal_enabled (bool, optional)
• timestamp_granularities (array<string>, optional): e.g. sentence.
• async (bool, optional): default false for sync models, true for qwen3-asr-flash-filetrans.

Response

• text (string): normalized transcript text.
• task_id (string, optional): present for async submission.
• status (string): SUCCEEDED or submission status.
• raw (object): original API response.

Quick start (official HTTP API)

Sync transcription (OpenAI-compatible protocol):

curl -sS --location 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions' \
  --header "Authorization: Bearer $DASHSCOPE_API_KEY" \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "qwen3-asr-flash",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "input_audio",
            "input_audio": {
              "data": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
            }
          }
        ]
      }
    ],
    "stream": false,
    "asr_options": {
      "enable_itn": false
    }
  }'

Async long-file transcription (DashScope protocol):

curl -sS --location 'https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription' \
  --header "Authorization: Bearer $DASHSCOPE_API_KEY" \
  --header 'X-DashScope-Async: enable' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "qwen3-asr-flash-filetrans",
    "input": {
      "file_url": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
    }
  }'

Poll task result:

curl -sS --location "https://dashscope.aliyuncs.com/api/v1/tasks/<task_id>" \
  --header "Authorization: Bearer $DASHSCOPE_API_KEY"

Local helper script

Use the bundled script for URL/local-file input and optional async polling:

python skills/ai/audio/alicloud-ai-audio-asr/scripts/transcribe_audio.py \
  --audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
  --model qwen3-asr-flash \
  --language-hints zh,en \
  --print-response

Long-file mode:

python skills/ai/audio/alicloud-ai-audio-asr/scripts/transcribe_audio.py \
  --audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
  --model qwen3-asr-flash-filetrans \
  --async \
  --wait

Operational guidance

• For local files, use input_audio.data (data URI) when direct URL is unavailable.
• Keep language_hints minimal to reduce recognition ambiguity.
• For async tasks, use 5-20s polling interval with max retry guard.
• Save normalized outputs under output/alicloud-ai-audio-asr/transcripts/.

Output location

• Default output: output/alicloud-ai-audio-asr/transcripts/
• Override base dir with OUTPUT_DIR.

Workflow

1. Confirm user intent, region, identifiers, and whether the operation is read-only or mutating.
2. Run one minimal read-only query first to verify connectivity and permissions.
3. Execute the target operation with explicit parameters and bounded scope.
4. Verify results and save output/evidence files.

References

• references/api_reference.md
• references/sources.md
• Realtime synthesis is provided by skills/ai/audio/alicloud-ai-audio-tts-realtime/.