Might not be the cheapest option, but this has the functionality you need.
First a get a silicon photodiode, http://www.osioptoelectronics.com/standard-products/silicon-photodiodes.aspx
Then get a blue light pass filter, http://opticalfiltershop.com/product-category/edge-filter/short-wave-pass-filters/
Now you make a circuit with a 1k resistor and connect this to the circuit in series with the photodiode. You can measure the voltage drop across this resistor and because you know the resistance you can calculate the current from the voltage drop.
Alternatively, you could make or purchase a nice photodiode (transconductance) amplifier if accuracy is important,
http://www.ti.com/lit/an/sboa061/sboa061.pdf
Filters that pass blue light but block other wavelengths are hard to find. Another option would be to disperse the light over different angular ranges, for example shining the light on to a prism or CD. Then only putting the diode where you see blue light. This would be similar to making your own monochromator.