This is probably better as a comment, but certainly too long.
I honestly think what you are looking for doesn't exist. Computer science, especially at the undergraduate level, has really become a vocational discipline. Most (undergraduate) students go into it for the purpose of getting software engineering jobs. The theoretical hurdles are usually regarded with some disdain (proofs tend to make folks cringe).
As a result, (undergraduate) CS texts have become more relaxed on the rigor. The Ullman-Hopcroft Automata Theory texts illustrate this point well. The first edition was far more rigorous and comprehensive than the third edition, which includes applications such as parsing HTML. A common criticism is that "all the good stuff" was taken out in writing the third edition.
Algorithms and complexity also go hand-in-hand. For decidable problems, the brute force and ignorance approach is usually enough to obtain a solution. For mathematicians, this may be enough for proofs (e.g., constructive proofs in graph theory). Computer scientists demand more in terms of algorithm efficiency. From a practical perspective, we want a computer program to be able to spit out a solution. From a theoretical perspective, demonstrating an algorithm correctly solves the problem within certain resource bounds enables us to classify computational problems into complexity classes.
Regarding references, I might suggest the following:
-Kleinberg and Tardos (https://www.amazon.com/Algorithm-Design-Jon-Kleinberg/dp/0321295358/). Both authors are top notch theorists. (Tardos was actually a Babai student as an undergrad, and Babai is certainly a stickler for the details.) The comments on Amazon suggest a great deal of mathematical rigor and emphasis on good algorithm design techniques. The table of contents looks solid to me.
-Sedgewick (https://www.amazon.com/Algorithms-4th-Robert-Sedgewick/dp/032157351X). A friend of mine went through Sedgewick's Coursera course and had excellent things to say. Based on his recommendation, Sedgewick may be worth considering. I'm not sure if it has the desired level of rigor.
More advanced texts on randomized algorithms or approximation algorithms may also have the desired level of rigor. Again, these are more advanced techniques, so perhaps not desirable as a starting point.
I hope this helps!