Summary
I checked that the sign of the Pcal inverse actuation filters is correct by injecting a known sine wave into the TRANSIENT filter bank and measuring at TRANSIENT_OUT and also the calibrated Tx PD signal, H1:CAL-PCALX_TX_PD_OUT_DQ. The signals are consistent and no minus sign is missing.

Details
To check the sign consistency, I made a 100 Hz sine wave with zero phase in Matlab and saved to an ASCII text file. Then I used awgstream to inject the waveform into H1:CAL-PINJX_TRANSIENT_EXC. I wanted to make sure that the output of H1:CAL-PCALX_TX_PD_OUT_DQ was correctly calibrated (not whitened) so I turned on the [:1,1] filter module, FM7. I then injected the waveform using awgstream and then measured the output in DTT, using a bandpass filter to see the 100 Hz signal.

I expect the Tx PD signal to be in phase with the injected waveform, while the TRANSIENT_OUT signal should be 180 degrees out of phase. The attached plot shows the output of H1:CAL-PCALX_TX_PD_OUT_DQ and H1:CAL-PINJX_TRANSIENT_OUT_DQ. The two signals are indeed 180 degrees out of phase, with the Tx PD output rising on the integer second as it should be, while the TRANSIENT_OUT is falling on the integer second (with some small delay, ~240 usec).

Thus, I conclude the sign of the inverse actuation filter (and, by extension, the actuation functions provided for the CW injections) is correct. If there is a sign flip, perhaps it is on the output side? This needs to be investigated.