The following is an example of script for clumping, you could adjust the parameters as needed:

plink --bfile  1000G_EUR \                                ### The LD reference panel
         --clump MetaGWAS.assoc \                         ### Your summary stats file
         --clump-p1 5e-6 \                                ### Significance threshold for index SNPs
         --clump-p2 0.05 \                                ### Secondary significance threshold for clumped SNPs
         --clump-r2 0.05 \                                ### LD threshold for clumping
        --out MetaGWAS_CAD_clumped_r0.05

Adding to this in case anyone else comes across it:

plink --bfile ref_panel \
    --clump your_sumstats \
    --clump-p1 5e-8 \
    --clump-p2 0.05 \
    --clump-r2 0.1 \
    --clump-kb 500 \
    --out output_file_name

In contrast to the above example:
p1: I've used the commonly accepted genome-wide significance threshold. You can use a less stringent threshold and then inspect and filter your clumped output accordingly.
r2: The default LD threshold for clumping is 0.5 but this is high. I tend to use 0.1 to be more conservative, the above 0.05 is even more conservative.
kb: This is missing above. This is the 'window' size, the genetic distance/radius that you are clumping within. The default 250kb is low. I would use at least 500kb but you often see 2Mb or 3Mb (i.e., setting the kb value to 2000 or 3000) to cover more extensive LD regions.

https://www.cog-genomics.org/plink/1.9/postproc