I'm on the team behind TAVEXA MINI (Android, package com.tavexa.mini). This category is full of fake apps, so we'd welcome scrutiny from people here. What it does: one-tap timed sessions calculated on our server, crews and weekly leagues. It shows optional rewarded ads (AdMob). What it doesn't do: n […]
Hi TavexaDev. The short version:
nothing in the disclosed permission set is an obvious malware marker, but a permission list and developer description are not enough to certify the app as trustworthy. The clearest description at present is an
ad-supported idle game whose server credits centrally controlled points—not cryptocurrency mining.
Permission observations
| Item | Assessment |
|---|
| Internet/network state/notifications/vibration | Expected for this design. |
| Boot completed | Plausible for restoring reminders, provided it does only that. |
| Exact alarms | Questionable. A session reminder normally does not require second-level precision, and gameplay is server-side anyway. Please identify whether this is SCHEDULE_EXACT_ALARM or USE_EXACT_ALARM and explain why an inexact alarm or push message is insufficient. |
| Foreground service/wake lock | Common in messaging libraries, but the release manifest should identify the actual service and foreground-service type. If it is never used, remove the declaration rather than blaming the library. |
| Biometric/fingerprint | These permissions let the app invoke Android authentication; they do not provide fingerprint images or templates. Not concerning by themselves. |
| C2DM | The old-looking name commonly remains in Firebase Cloud Messaging declarations. |
| Advertising ID/install referrer | Not dangerous permissions, but privacy-relevant. They support advertising, reward attribution and campaign measurement. |
On Android, users install one merged manifest. A dependency does not receive diplomatic immunity: permissions introduced by an SDK remain the developer’s responsibility.
Your privacy policy is materially more informative than the opening post. It says the service processes account and gameplay records, push tokens, a Play Integrity verdict with a random installation identifier, IP addresses briefly for abuse controls, AdMob identifiers/interactions, diagnostics, and optional community or Telegram information. That is not necessarily improper, but it illustrates an important distinction:
“no contacts/location permission” does not mean “little or no data processing.” The Play Data Safety declaration should match all of those flows, including SDK behaviour.
The larger trust problem is the presentation
The screenshots use “mining network,” permanent early-miner bonuses, limited Genesis spots and “mine faster” language. That does not prove fraud, but it borrows nearly the complete wardrobe of speculative crypto schemes while simultaneously saying there is no cryptocurrency. A cautious user will reasonably wonder whether a future token or withdrawal promise is waiting offstage.
If TVX is permanently just a game score, say prominently and consistently:
- TVX cannot be withdrawn, transferred, sold or converted to money or cryptocurrency.
- Starting a session performs no computation on the device.
- Rewards are database entries controlled by TAVEXA and may be changed under the game rules.
- Rewarded ads are optional and refusing them does not prevent ordinary participation.
The screenshots also show a
Shop, while the privacy policy refers to “in-app store purchases.” If these are virtual items purchased only with earned TVX, call them
TVX-only virtual purchases. If any real-money purchase exists, “nothing to buy to take part” is technically defensible but less clear than simply disclosing it.
Calling this an “idle social game” would be more accurate than calling it a mining network. Marketing may prefer the latter; trust prefers the former.
What would make this meaningfully auditable
Please publish, for each release:
- The merged release manifest and manifest-merger report—not only the permissions listed in source files.
- Version code/name, signing-certificate SHA-256 fingerprint and release date.
- An SDK inventory with exact versions: AdMob, Firebase components, sign-in, Supabase and anything else.
- The expected network domains and the purpose of each.
- A short explanation for every boot receiver, service and foreground-service declaration.
- Confirmation that denying notifications, exact-alarm access, ad consent or biometric use does not break the basic game.
- Independent static and dynamic results, including CPU/battery behaviour while a session supposedly “mines” with the app closed.
A particularly useful test would be to start a four-hour session, close the app and verify that no persistent service, sustained CPU use or unexplained traffic remains. That would support the no-on-device-mining claim. It still would not independently verify the backend economy: “server authoritative” improves anti-cheat, but it also means users must trust your server and your database.
Provisional verdict: no malware evidence is established by what is currently shown, and most listed permissions are explainable. The exact-alarm and foreground-service declarations deserve closer examination. The biggest obstacle to trust is not a scary permission—it is crypto-style scarcity marketing for a centrally operated, non-redeemable game counter.
Sources