JWST's line-of-sight stability, based on actual performance, is covered in this article.

Guiding precision and pointing stability

For fixed targets, pointing stability is evaluated as the rms error in the guide star position in a 15 s interval, compared to the mean position over a 10,000 s observation. 

Line-of-sight (LOS) pointing stability under fine guiding control is excellent. The Fine Guiding Sensor (FGS) sensing precision is parameterized as the noise equivalent angle (NEA, an equivalent jitter angle based on centroid position and SNR), and is usually ~1 mas (1-σ per axis), symmetric. The LOS jitter seen in the science instruments is similarly good, and high frequency sampling of the jitter in a NIRCam small subarray is the same, ~1 mas (1-σ per axis). Although a drift in observatory roll could theoretically generate image drift in a long duration observation, measurements during commissioning indicated such drifts are negligible. 

For Solar System (i.e., moving) targets, the line-of-sight pointing stability is evaluated as the rms value over a 1,000s observation for a linear motion of 3.0 mas/s, and is estimated to be 6.2 to 6.7 mas (1-σ per axis, depending on the science instrument).

Measuring jitter

Jitter is measured at the position of the guide star in the FGS, then mapped to the differential distortion compensation field point specified in the engineering telemetry. Jitter data is available in the "Pointing " table extension for Track and Fine Guide data FITS files. 

Figure 1 shows a typical jitter profile. This particular dataset was obtained from the first jitter analysis performed during commissioning (PID 1163, observation 2, reference JWST-STScI-008271).

References

Gardner, J. P. et al. 2006, Space Sci.Rev., 123, 485 
The James Webb Space Telescope

Hartig, G. F., & Lallo, M. 2022, JWST-STScI-008271 
JWST Line-of-Sight Jitter Measurement during Commissioning