If you do not mind downloading the human reference genome (~900MB), you can:

wget --no-check-certificate https://raw.github.com/lh3/misc/master/seq/seqtk/seqtk.c
wget --no-check-certificate https://raw.github.com/lh3/misc/master/seq/seqtk/khash.h
wget --no-check-certificate https://raw.github.com/lh3/misc/master/seq/seqtk/kseq.h
wget ftp://ftp-trace.ncbi.nih.gov/1000genomes/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz
gcc -g -Wall -O2 seqtk.c -lz -o seqtk
sed s,^chr,, your.list | awk -v OFS="\t" '{print $1,$2-26,$2+25}' > snps.bed
./seqtk subseq -t hs37d5.fa.gz snps.bed > snps.seq
paste your.list snps.seq > snps.final.txt

As to UCSC, I do know where the reference sequence is kept in MySQL, but for such a task, it is frequently simpler to do locally. The human reference genome is pretty small in comparison to other data sets. BTW, if you want to get the a sequence in a couple of regions, a more convenient way is:

samtools faidx ftp://ftp-trace.ncbi.nih.gov/1000genomes/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz 1:1,000,000-1,010,000 2:2,000,000-2,010,000

But this method becomes inefficient when you want to get thousands of regions.

for this purpose I used to suggest to my lab colleages this script I found years ago, which queries EMBL and GenBank, from the very interesting bioperl repository. but as Jeremy has pointed out, I see great solutions on this post which I don't see why they wouldn't suit your needs.