I'm reading Volume 1 of Quantum Fields and Strings: A Course for Mathematicians.
Let $V$ be a Minkowski vector space of signature $(1,d-1)$, and let $\mathrm{Spin}(V) \curvearrowright S$ be an irreducible real spin representation.
Fix the gamma map $\Gamma: S^* \otimes S^* \to V$.
The super Lie algebra $\mathcal{L}$ is defined as $V \oplus \Pi S^*$ as a vector space, where $V$ is central, and
$$ [s,t] = -2 \Gamma(s,t) $$
for $s,t \in \Pi S^*$. Here $\Pi$ denotes odd parity.
Let $\exp(\mathcal{L})$ be the Lie supergroup with Lie algebra $\mathcal{L}$.
Fix bases $\{e_\mu\} \subset V$ and $\{f_a\} \subset \Pi S^*$.
I want to show that
$$ D_a = \partial_a - \theta^b \Gamma_{ab}^\mu \partial_\mu $$
is the left-invariant vector field on $\exp(\mathcal{L})$ corresponding to $f_a \in \mathcal{L}$, but I have encountered a sign problem.
From the BCH formula, I think the group law is
$$ (x^\mu, \theta^a) = (x_1^\mu + x_2^\mu - \Gamma_{bc}^\mu\theta^b_1 \theta^c_2, \theta^a_1 + \theta^a_2) $$
Let $m: \exp(\mathcal{L}) \times \exp(\mathcal{L}) \to \exp(\mathcal{L})$ be the group multiplication. The left-invariance condition means
$$ \begin{aligned} (0, D_a)m^* x^\mu &= m^*(D_a x^\mu), \\ (0, D_a)m^* \theta^b &= m^*(D_a \theta^b). \end{aligned} $$
The second equation poses no problem. However,
$$ \begin{aligned} (0, D_a)m^* x^\mu &= (\partial_{a, 2} - \theta^b_2 \Gamma_{ab}^\nu \partial_{\nu, 2})(x_1^\mu + x_2^\mu - \Gamma_{bc}^\mu\theta^b_1 \theta^c_2) \\ &= -\Gamma_{bc}^\mu\partial_{a, 2}(\theta^b_1 \theta^c_2) - \theta^b_2 \Gamma_{ab}^\mu \\ &= \Gamma_{bc}^\mu\theta^b_1\partial_{a, 2}(\theta^c_2) - \theta^b_2 \Gamma_{ab}^\mu \\ &= \Gamma_{ba}^\mu\theta^b_1 - \theta^b_2 \Gamma_{ab}^\mu \\ &= \theta^b_1\Gamma_{ab}^\mu - \theta^b_2 \Gamma_{ab}^\mu, \\ m^*(D_a x^\mu) &= m^*(-\theta^b \Gamma_{ab}^\mu) = -(\theta_1^b + \theta_2^b) \Gamma_{ab}^\mu, \end{aligned} $$
which do not coincide.