A tiny Python script (only 30 loc) that, given a C file, generates documentation for it. The output is a markdown file, making it ideal for use with GitHub.

The full code is shown below:

import sys, re

class T:
    def __init__(self, text): self.curr = 0; self.txt = text.splitlines(True)
    def at(self, i): return self.txt[i] if i < len(self.txt) else None
    def peek(self): return self.at(self.curr)
    def next(self): self.curr += 1; return self.at(self.curr - 1)

def d_start(l): return re.search(r"^\/\*\*\s*(\w*)", l)
def d_end(l): return re.search(r"^ \*\/", l) if l else True
def d_line(l): return re.search(r"^ ?\*? ?(.*)", l).groups()[0] + "\n"

def d_iter(txt):
    while not d_end(l := txt.next()): yield d_line(l)

def s_iter(txt):
    while txt.peek() and not d_start(txt.peek()): yield txt.next()

def t_iter(txt):
     while l := txt.next():
        if not d_start(l): continue
        match d_start(l).groups()[0]:
            case "code":
                d, s = "".join(d_iter(txt)), "".join(s_iter(txt)).strip()
                yield "\n```c\n" + s + "\n```\n\n" + d + "\n----\n"
            case "text": yield "\n" + "".join(d_iter(txt))
            case other: raise ValueError("unknown doc type found: " + other)

assert len(sys.argv) == 3, "provide 2 arguments"
open(sys.argv[2], "w").write("".join(t_iter(T(open(sys.argv[1], "r").read()))))

microcompute uses this tool to generate documentation. The output can be seen here.