transcribe

Speech-to-text via SkillBoss API Hub (STT, powered by Whisper and more).

Security Scan

The skill is internally consistent with its stated purpose (sending local audio to SkillBoss for STT) and only requests a single API key; the primary risk is accidental exposure of sensitive audio when files are uploaded to the external service.

Purpose & Capability

Name/description (speech-to-text via SkillBoss) align with the declared requirement (SKILLBOSS_API_KEY) and the runtime instructions which call SkillBoss's /v1/pilot endpoint. The requested environment variable is appropriate for calling a third‑party API.

Instruction Scope

SKILL.md instructs the agent to read local audio files (open('/path/audio.mp3')), base64-encode them, and POST to an external API. This is expected for a transcribe skill, but it does mean any audio the agent is pointed to will be uploaded to SkillBoss — a possible privacy/exfiltration concern if sensitive audio is used.

Install Mechanism

Instruction-only skill with no install spec or bundled code; nothing is written to disk by the skill itself, lowering supply-chain risk.

Credentials

Only a single API key (SKILLBOSS_API_KEY) is required, which is proportional to calling a third-party STT API. The key grants access to the external service and could incur usage/billing; no unrelated credentials are requested.

Persistence & Privilege

The skill is not marked always:true and has default invocation settings. It does not request system-wide configuration or modify other skills. No elevated persistence is requested.

Assessment

This skill appears to do what it says: it reads local audio files and uploads them to SkillBoss for transcription using the SKILLBOSS_API_KEY. Before installing, confirm you trust api.skillbossai.com and its privacy/billing practices, avoid transcribing highly sensitive audio unless you accept uploading it to the service, store the API key securely (and rotate it if compromised), and limit the file paths the agent is allowed to access. If you need on-device transcription or tighter data control, consider a local/offline STT solution instead.

Whisper STT via SkillBoss API Hub

Use SkillBoss API Hub's /v1/pilot to transcribe audio (STT), powered by OpenAI Whisper and other speech recognition models.

Quick start (Python)

import requests, base64, os

SKILLBOSS_API_KEY = os.environ["SKILLBOSS_API_KEY"]
API_BASE = "https://api.skillbossai.com/v1"

def pilot(body: dict) -> dict:
    r = requests.post(
        f"{API_BASE}/pilot",
        headers={"Authorization": f"Bearer {SKILLBOSS_API_KEY}", "Content-Type": "application/json"},
        json=body,
        timeout=60,
    )
    return r.json()

# Transcribe audio file
audio_b64 = base64.b64encode(open("/path/audio.mp3", "rb").read()).decode()
result = pilot({"type": "stt", "inputs": {"audio_data": audio_b64, "filename": "audio.mp3"}})
text = result["result"]["text"]
print(text)

# Translate audio to English
result = pilot({"type": "stt", "inputs": {"audio_data": audio_b64, "filename": "audio.m4a", "task": "translate"}})
text = result["result"]["text"]
print(text)

Notes

• No local model download required; SkillBoss API Hub automatically routes to the best STT model.
• SKILLBOSS_API_KEY environment variable required.
• Response text is at result["result"]["text"].