SenseAudio-ASR

Build and troubleshoot SenseAudio speech recognition integrations, including HTTP transcription (`/v1/audio/transcriptions`), realtime WebSocket ASR (`/ws/v1...

Security Scan

The skill's requested secrets, instructions, and behavior are consistent with a SenseAudio ASR integration and do not ask for unrelated privileges or risky installs.

Purpose & Capability

Name/description match the requested asset (SENSEAUDIO_API_KEY) and the documented endpoints/models. There are no unrelated env vars or binaries requested that would be inconsistent with an ASR integration. (Note: the skill source is listed as unknown — that affects trust but not internal coherence.)

Instruction Scope

SKILL.md and references/asr.md limit actions to building requests against SenseAudio endpoints, handling responses, and treating identified operational fields as sensitive. The instructions do not direct the agent to read unrelated files, system credentials, or to exfiltrate data to third-party endpoints.

Install Mechanism

Instruction-only skill with no install spec and no code files, so nothing is written to disk or downloaded during install — lowest-risk install surface.

Credentials

Only a single API key (SENSEAUDIO_API_KEY) is required and declared as primaryEnv, which is appropriate for a third‑party ASR integration. The documentation explicitly warns against logging or embedding the key.

Persistence & Privilege

always is false and the skill does not request system-wide configuration or modification of other skills. Autonomous invocation is allowed (platform default) but it is not combined with broad credentials or elevated privileges.

Assessment

This skill appears coherent for integrating with SenseAudio: it only needs your SENSEAUDIO_API_KEY and documents how to use it safely. Before installing, verify the provider (senseaudio.cn) and prefer using a scoped/test API key with minimal permissions and short lifetime. Do not reuse high-privilege keys; enable rotation and monitoring. If you need stronger assurance, ask the publisher for a canonical source repository or signed documentation since the skill source is marked unknown.

Use this skill for all SenseAudio speech recognition tasks.

Credential source: read the API key from SENSEAUDIO_API_KEY and send it only in the Authorization: Bearer ... header. Do not place API keys in query parameters, logs, transcripts, or saved examples.

Read First

• references/asr.md

Workflow

1. Pick recognition mode:

• HTTP file transcription for offline audio.
• WebSocket for realtime streaming microphone/audio chunks.
• Audio analysis for noise and quality checks before recognition.
• Records query for recent recognition history lookup.

2. Choose model by feature needs:

• Lite for low-cost basic transcription.
• ASR for streaming, translation, diarization, sentiment, and timestamps.
• Pro when diarization plus explicit max_speakers control is needed.
• DeepThink for streaming, translation, and intelligent editing; do not send language, diarization, sentiment, timestamps, ITN, or punctuation controls.

3. Build minimal request:

• Required auth, file/audio format, model.
• Add optional controls only when needed.
• Keep uploaded files at or below 10MB; split longer audio before sending.

4. Validate compatibility:

• Check model-parameter support before sending.
• Enforce WS pcm / 16000Hz / mono requirements.
• For HTTP stream=true, expect SSE text deltas only, not structured verbose fields.

5. Parse robustly:

• Handle JSON/text/verbose/SSE forms.
• Handle WS terminal events and failures.
• Treat returned audio URLs, api_key, session_id, and trace_id as sensitive operational data.