Attack - KeeShare attachments can be inferred because of attachment de-duplication.
Solution - Prevent de-duplication of normal database entry attachments with those entry attachments synchronized/associated with a KeeShare database. This is done using the KeeShare database UUID injected into the hash calculation of the attachment prior to de-dupe. The attachments themselves are not modified in any way.
--------
Attack - Side channel byte-by-byte inference due to compression de-duplication of data between a KeeShare database and it's parent.
Solution - Generate a random array between 64 and 512 bytes, convert to hex, and store in the database custom data.
--------
Attack vector assumptions:
1. Compression is enabled
2. The attacker has access to a KeeShare database actively syncing with the victim's database
3. The victim's database is unlocked and syncing
4. The attacker can see the exact size of the victim's database after saving, and syncing, the KeeShare database
Thank you to Andrés Fábrega from Cornell University for theorizing and informing us of this attack vector.
* Closes#7545 - Support 1Password 1PUX import format based on https://support.1password.com/1pux-format/
* Closes#8367 - Support Bitwarden JSON import format (both unencrypted and encrypted) based on https://bitwarden.com/help/encrypted-export/
* Fixes#9577 - OPVault import when fields have the same name or type
* Introduce the import wizard to handle all import tasks (CSV, KDBX1, OPVault, 1PUX, JSON)
* Clean up CSV parser code to make it much more efficient and easier to read
* Combine all importer tests (except CSV) into one test file
* Include check for group as recycle bin directly into the Group::isRecycled() function
* Return the original root group from Database::setRootGroup(...) to force memory management transfer
* Previously our base style sheet added roughly 20px of margin to the top and bottom of all QGroupBox. This caused visual errors where that margin was not needed/desired.
* Transferred padding to the specific layouts instead where it belongs.
Refactors the Passkey implementation to include more checks and a structure that is more aligned with the official specification.
Notable changes:
- _BrowserService_ no longer does the checks by itself. A new class _BrowserPasskeysClient_ constructs the relevant objects, acting as a client. _BrowserService_ only acts as a bridge between the client and _BrowserPasskeys_ (authenticator) and calls the relevant popups for user interaction.
- A new helper class _PasskeyUtils_ includes the actual checks and parses the objects.
- _BrowserPasskeys_ is pretty much intact, but some functions have been moved to PasskeyUtils.
- Fixes Ed25519 encoding in _BrowserCBOR_.
- Adds new error messages.
- User confirmation for Passkey retrieval is also asked even if `discouraged` is used. This goes against the specification, but currently there's no other way to verify the user.
- `cross-platform` is also accepted for compatibility. This could be removed if there's a potential issue with it.
- Extension data is now handled correctly during Authentication.
- Allowed and excluded credentials are now handled correctly.
- `KPEX_PASSKEY_GENERATED_USER_ID` is renamed to `KPEX_PASSKEY_CREDENTIAL_ID`
- Adds a new option "Allow localhost with Passkeys" to Browser Integration -> Advanced tab. By default it's not allowed to access HTTP sites, but `http://localhost` can be allowed for debugging and testing purposes for local servers.
- Add tag `Passkey` to a Passkey entry, or an entry with an imported Passkey.
Fixes#10287.
* Entry: re-parent before adding to new group
Adding the Entry to the Group will emit signals about the action.
Present the object with the correct parent already.
* fdosecrets: Item::Create() can fail
If an entry cannot be registered on DBus, Item::Create() will return a
nullptr. Basically, this can only happen if there is already an item
with the same UUID in the collection. The only viable option here is to
ignore the new entry.
* Merger: prevent duplicate entry when merging histories
If the source entry is newer, a copy of the entry is made. But before
moving the merged entry to the target group, it must be removed.
Otherwise there will be briefly two entries with the same UUID
in the same group/database.
Even though this is only the case during the transaction, it can still
be observed because the operations emit signals. A notable problem is
the fdosecrets feature that relies on the uniqueness of the UUID or will
otherwise run into problems because the UUID is used as part of the DBus
path.
* Fix#9847 - don't provide TOTP values if settings are blank or completely wrong
* Fix#6838 - don't reset the ui when creating a new entry and applying TOTP to it
* Move totp source into the core folder
Botan::secure_scrub_memory -> defined in mem_ops.h
Botan::secure_vector -> defined in secmem.h
The reason only including secmem.h worked in previous (<3.0) versions of
Botan was because secmem.h included mem_ops.h. This is no longer the
case since commit
randombit/botan@49dbbcb2bf (2023-10-11;
"Split out allocator helpers to allocator.h")
Fixes#10038
The recent debug library is called botan-3.lib, which is unable to be found by this since BOTAN_NAMES_DEBUG does not contain the keyword "botan-3". This commit adds that keyword.
* Update release-tool.ps1
`Get-CimInstance MSFT_VSInstance` does not always work as it uses the default namespace root/cimv2. Specifying `Get-CimInstance MSFT_VSInstance -Namespace root/cimv2/vs` allows it to work.
Co-authored-by: Jonathan White <support@dmapps.us>
