If you don't want to size your pond for the off-site flow, then you have to divert it around the detention pond.  The detention pond doesn't distinguish where the water came from when it's filling up.

If I size my outfall structure to handle the total offsite flow how does the detention pond gain in size?  That is the question I cannot answer.

You're probably obligated to maintain the pre-developed flow rate, which is why you're installing a detention basin.  (You're slowing the water down so that it only leaves the site as fast as it was previously.)

If you don't increase the size of your detention basin, chances are, your outfall pipe will have to be so big that in frequent storm events (eg. 2-year storm) you're exceeding the pre-developed flow rate.  If you keep your pipe small enough to meet the flow rate criterion in frequent storm events, your larger storm events (25-year, 50-year, 100-year, etc.) will likely overtop your detention basin.  When this happens, it usually happens close to the peak flow period on your inflow hydrograph, which means that most of the peak is going straight over the top of the detention basin and you're not detaining it at all.

Since the offsite flow is an existing condition you should not have to detain this flow.  Your basin design should attenuate the increase of flow due to your improvements to match the existing condition.  You have to incorporate the offsite flow into your model if it enters your basin.  The outflow of your basin should match that of your existing condition.  Remember that this is an iterative process that will take time to finalize.  I assume you have initially sized your basin and have storage routed your hydrogragh through it.  Did the basin attenuate your flow rate?

If you have not initially sized your basin, my preferred method is to plot the inflow hydrogragh and draw a line from the baseflow/zero point to the preferred outflow.  The area below the hydrogragh and above the line is the volume you need for your basin, increase by 10%.  This is only a starting point.  The storage volume must be tweaked.  I have not used the modified rational method for deten basin design.  I always revert back to my preferred method.

Be careful ! "Modified Rational Method " may mean different things in different areas of the country.  In many places in California it is very different than in say, Washington State.

If you are using the Rational method you won't have a hydrograph.  You can make one up but the method does not yield a hydrograph; only a peak flow. For this reason, many engineers prefer other methods such as TR-55 or the Santa Barbara Urban Hydrograph, or even the National Flood Frequency (NFF) program.  Reginal Regression equations may also be used and are pretty conservative in most cases.

For these, and many other reasons, regional detention makes more sense than "local", i.e. on site detention.

good luck

I would analyze the pre-development flow very carefully in this case.

To what extent can you devert the existing offsite flow?  Is it economical?  What will happen to that flow during the 100 year event?  Perhaps you should perform two or three preliminary designs, one considering the offsite flow, one without the flow and maybe a third using a percentage of the flow.   When in doubt, it is better to be on the conservative side with your design.  Document your your findings and present them to the city engineer.  Remember it is better to work with the city engineering department than against them.  (I know, they can be unreasonable at times.)

It is difference between the Pre and Post-development flow that you are concerned about detaining, so I would look at it from several angles.  A lot of engineering "judgement" is required in these cases.