README.md: fix link and subdir prefixes

Signed-off-by: Tim Janik <timj@gnu.org>

README.md

@@ -2,14 +2,14 @@
 # SCL
 
 The Small/simple C/C++ Library (SCL) provides various algorithms in a portable way, with little to no external dependencies.
-The library is written in C++17 and uses the [MPL-2.0 License](http://mozilla.org/MPL/2.0) or [Public Domain](https://unlicense.org/UNLICENSE) dedications.
+The library is written in C++17 and uses the [MPL-2.0 License](http://mozilla.org/MPL/2.0) or [Public Domain](https://unlicense.org/) dedications.
 
 The subdirectories are structured as follows:
 
-* `scl/shishua` contains a Public Domain Shishua PRNG implementation for generic CPUs (ALU), the SSE2/3 and AVX2 instruction set variants.
+* `shishua/` contains a Public Domain Shishua PRNG implementation for generic CPUs (ALU), the SSE2/3 and AVX2 instruction set variants.
   This is currently the fastest known PRNG implementation, it passes PractRand at 32TB.
-* `scl/chacha` contains a Public Domain ChaCha block cipher implementation for generic CPUs (ALU), the SSE2/3 and AVX2 instruction set variants.
+* `chacha/` contains a Public Domain ChaCha block cipher implementation for generic CPUs (ALU), the SSE2/3 and AVX2 instruction set variants.
   This is probably the fastest CSPRNG implementation, it passes PractRand at 32TB.
-* `scl/keccak` contains a Public Domain Keccak PRNG implementation for generic CPUs (ALU) and system entropy gathering code according to the
+* `keccak/` contains a Public Domain Keccak PRNG implementation for generic CPUs (ALU) and system entropy gathering code according to the
   "Welcome to the Entropics: Boot-Time Entropy in Embedded Devices" paper.
   Using Keccak allows for a 1600 bit CSPRNG entropy pool, it passes PractRand at 32TB.