Okay, I got vecscreen to work. The problem was that the app wasn't included in the FTP files that I downloaded from NCBI. I used subversion to get all the code and was able to find and build vecscreen. The text output can be used to clean the sequences. This could be done with Python and BioPython.
Here is what I finally did to get vecscreen to compile. As I mentioned, for some reason it wasn't in the tarball from the FTP site, so I had to check out with subversion (svn)
NCBI toolbox users manual for building: http://www.ncbi.nlm.nih.gov/books/NBK7167/
With Linux ...
Make sure G++ is installed (could be different for different platforms), etc.
Use the following command to get the source:
svn co http://anonsvn.ncbi.nlm.nih.gov/repos/v1/trunk/c++
From the compilers directory, do ./GCC.sh (this is different for different platforms) This step could be unnecessary
From the top-level directory of the checked out files, do
make -f Makefile.flat $PROJECT_NAME (i.e. app/vecscreen/)
Also need app/blast/ and app/blastdb/
Downloaded UniVec_Core fasta file from ftp://ftp.ncbi.nih.gov/pub/UniVec/ (This has only non-mammalian vectors)
Make local copy of UniVec_Core database in the GCC444-Debug/bin directory with the command:
./makeblastdb -in UniVec_Core -dbtype nucl -out UniVec_Core.db
Use the vecscreen command (found in GCC444-Debug/bin/)
./vecscreen -db UniVec_Core.db -query $fasta_file -out $vecscreen_outfile -outfmt 0 -text_output