Cartan's first structural equation

(@lee2006riemannian exercise 4-5 page 64)
Let be a linear connection on M, let {ei} be a local frame on some open subset UM, and let {ωi} be the dual coframe. Denote by Tijk the structure coefficients.
We know that there is a uniquely determined matrix of 1-forms Θji on U, called the connection 1-forms for this frame, such that

Xei=XΘijej

for all XTM.

Theorem (Cartan’s first structure equation)

dωk=ωiΘik+τk,

where {τ1,...,τn} are the torsion 2-forms, defined in terms of the torsion and the frame {ei} by

τ(X,Y)=τj(X,Y)ej.


Proof
We take e1,e2 without lost of generality, and observe that the torsion satisfies

τ(e1,e2)=e1e2e2e1[e1,e2]

The kth component will be

τk(e1,e2)=e1Θ2ke2Θ1k+T12k

But also dωk(e1,e2)=T12k, since dωk=i<jTijkωiωj. It only rests to show that

ωiΘik(e1,e2)=e2Θ1ke1Θ2k

and this is true, since

ωiΘik(e1,e2)=idet(e1ωie1Θike2ωie2Θik)=e2Θike1Θik.

Comment: I think that Cartan's first structural equation is used sometimes to recover a connection expressed in a particular frame/coframe if we know the torsion. Mainly with the Levi-Civita connection of a Riemannian metric and orthonormal frames
In general, you need to know more information about the "relation" of the chosen frame with the connection. But in the particular case in which the frame is orthonormal, now this is enough data to recover the metric using the torsion (null torsion, for example). But anyway, if you assume orthornormality you can compute the metric, compute Christoffel symbols and then change to the desired frame, not needing Cartan's first structure equations... I guess they are used as a shortcut, but they are not indeed needed...

Related: Cartan's second structural equation