How to parse nested, mini grammar in string using Coco/R?

Question

I need help constructing the Coco/R grammar to match the following:

// Expects "1+2=3" (String concat. No embedded expression)
"1+2="+(1+2)

// Expects "1+2=3" (Embedded expression)
"1+2=%(1+2)"

// Expects "1+2=3" (unlikely, but still valid embedded expression)
"%(pow(0,1-1))+%(sqrt(4))=%(10/2-2)"

I'm trying to match embedded expressions in a string (within %())

Here's the grammar that I currently have.

It matches all 3 examples above. But now I have no idea how to continue.

COMPILER FooLang

CHARACTERS
    letter = 'a'..'z'.
    digit = '0'..'9'.
    strChar = ANY - '"'.

TOKENS
    ident = letter {letter}.
    number = digit {digit}.
    string = '"' {strChar} '"'.

PRODUCTIONS
    FooLang = expr.
    expr = equ.
    method = ident '(' [ expr {',' expr} ] ')'.
    equ = add { "==" add}.
    add = mul { ('+'|'-') mul }.
    mul = atom { ('*'|'/') atom }.
    atom = number | string | method | '(' expr ')'.

END FooLang.

The grammar within %() will be almost the same, but with some differences (ie, no string tokens)

Here's what I've found so far:

1. Island grammar antlr3
2. Handling Multiple Language Contexts by Subtractive Context Switching
3. Island Grammars
4. ANTLR 4 lexer tokens inside other tokens

It appears that I want Antlr's "lexer modes", but for Coco/R.