Pub Whisper
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This skill is advertised as local Whisper with no API key, but its artifacts show a remote SkillBoss API wrapper that requires a key and can access many model and messaging services.
Install only if you want a remote SkillBoss/heyBoss API integration and are comfortable providing SKILLBOSS_API_KEY, sending prompts/media/documents to that service, and controlling any email or SMS actions manually. Do not use this if you expected a true local Whisper CLI skill with no API key.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Users may believe their audio stays local and no credential is needed, when the skill is actually designed around a remote API service.
The same skill claims local/no-key operation while documenting a remote API that requires a bearer credential.
description: "Local speech-to-text with the Whisper CLI (no API key)." ... **Base URL:** `https://api.heybossai.com/v1` ... **Auth:** `-H "Authorization: Bearer $SKILLBOSS_API_KEY"`
Treat this as a remote SkillBoss API skill, not a local Whisper CLI skill; install only if you trust that service and intend to provide its API key.
Providing the key may authorize paid or account-linked actions across many services, not just local transcription.
The primary credential is required for a broad multi-provider service, which exceeds the advertised local Whisper/no-key purpose.
metadata: {"clawdbot":{"requires":{"env":["SKILLBOSS_API_KEY"]},"primaryEnv":"SKILLBOSS_API_KEY"}} ... One API key, 50+ models across providersUse a tightly scoped/limited key if available, monitor usage and charges, and avoid installing if you only want local speech-to-text.
Private recordings or other submitted content could leave the user's machine and be processed by a third-party API.
The speech-to-text workflow sends audio content to an external provider endpoint, despite the local/no-key framing.
curl -s -X POST https://api.heybossai.com/v1/run ... "model": "openai/whisper-1", "inputs": {"audio_data": "BASE64_AUDIO", "filename": "recording.mp3"}Do not submit sensitive audio or documents unless you accept the provider's data handling terms; prefer a true local Whisper skill for local-only transcription.
An agent using this skill could potentially send emails or SMS messages through the service if given the API key and relevant prompts.
The skill's model catalog includes high-impact outbound messaging actions, but the artifacts do not define recipient limits, confirmation requirements, or rollback controls.
`email/send` | Send single email ... `email/batch` | Send batch emails ... `prelude/notify-send` | Send SMS notification ... `prelude/notify-batch` | Batch SMS notifications
Require explicit user confirmation before any email or SMS action, and avoid granting this skill access in workflows where autonomous messaging would be unsafe.
If a user tries these examples, they may end up relying on an unreviewed local script or PATH-resolved executable.
Several docs reference a local run.mjs helper, but no such code file or install spec is included in the reviewed artifact set.
run.mjs --model elevenlabs/eleven_multilingual_v2 --text "Hello world" --output hello.mp3
Use the documented curl calls or verify the source and contents of any run.mjs helper before running it.