What about shred(1)?

That depends -- if only the file contents are precious, then shred will securely wipe the data blocks, since with the default mount options, file data doesn't go through the journal. If you're using data=journal mode, you can always write $JOURNAL_SIZE amount of data to fill up the journal thereby overwriting anything that was in it. If, however, the metadata was also precious, then yes, shred alone might not have been sufficient. Though it isn't difficult to change most of the fields in an inode/extent tree/attributes from userspace.