Hello,
I maintain VaultKeepR, a local-first, zero-knowledge password vault (Argon2id + XChaCha20-Poly1305, IPFS-based encrypted sync, wallet-signature login). It has been listed on AMO for X months, and version 1.8.2 was live without any issue.
Last week we submitted 1.8.3. It was our first version carrying the mandatory browser_specific_settings.gecko.data_collection_permissions key, where we declared “required”: [“none”]. A manual review then disabled the add-on under the Acceptable Use policy, clause “Deceptive or misleading”, with no further detail.
We believe the reviewer reconciled that declaration with three opt-in flows that do transmit data outside the extension:
- the email breach check sends the user’s addresses to leakcheck.io after an explicit consent dialog that names the service
- the Terms-of-Service analyzer fetches the public pages the user asks it to summarize, and does the analysis on-device
- enterprise SSO join verifies a short-lived capability id against our backend, which returns the verified work email
None of these are required for the vault to work, but a bare “none” with no optional types was inaccurate on our side. For the record, the bundles contain no analytics or telemetry SDK; strings like “mixpanel” or “fathom” that a source scan surfaces belong to the tracker-detection lists of the TOS analyzer, not to any integration.
My questions:
- Does this look like the kind of fix that gets an Acceptable Use decision reversed, or is that clause usually final?
- While the appeal is pending, is it acceptable (or counterproductive) to submit the corrected version as a new listing?
- Is there anything reviewers commonly expect in this situation that we may have missed?
Thanks for reading.