everyone has personality traits (and almost everyone has traits of personality disorder), but yes you often see this b.s. "personality traits" in people's charts. Absolutely avoid it if they do not meet criteria for personality disorder. You don't need to know a person for any length of time to diagnose personality disorder. I might see someone once and diagnose psychopathy, borderline personality disorder etc. a personality disorder diagnosis should ideally be based on a history including detailed premorbid personality history, and social and developmental history. Collateral information, and review of the medical chart and any other records is helpful. You should not base a diagnosis of personality disorder on a mental status examination alone as many mental disorders (depression, bipolar disorder, dementia, alcoholism, pathological gambling, autism, psychosis) may mimic personality disorder.
Personality disorder diagnosis rely on stability over time and range of events. Unfortunate, the stability, and reliability of personality disorder diagnoses is quite poor. People can meet criteria at one point in time, and not another. People can appear personality disorder in response to one crisis but not a similar one at a different point. That is why personality disorder experts suggest moving away from the personality disorder concept to a concept of impaired personality functioning. This accounts for the fact that someone might have very significant impairment in interpersonal, affect regulation, impulse control and so on at a particular point in time, without there being a better descriptor but does not represent a life long condition. Also cluster B personality disorders tend to get better with age. In contrast, cluster C personality disorders tend to become much worse with age.
The only exception to using the term "personality traits" is when you are diagnosing general personality disorder. If the patient meets criteria for general personality disorder and has say 4 criteria for borderline personality disorder, than a diagnosis of personality disorder with borderline personality traits can be made. Similarly, if a patient meets full criteria for one or more personality disorders, and has significant number of relevant criteria from another diagnosis but does not meet threshold then a diagnosis of say borderline personality disorder with narcissistic traits could be made. I would only do this if you can justify the relevance of the additional personality traits to the disorder. Particularly on the east coast it is most common for people to talk about borderline personality organization rather than "borderline personality traits". One proposal for new personality disorder criteria is to get rid of all the different personality disorder diagnoses (which are famously unreliable and lack validity) and replace it with simply "personality disorder" and depending on how many criteria they have (which corresponds with impairment and risk - for example those with 8 or more criteria with BPD will more likely than not commit suicide) stratify as "mild", "moderate" or "Severe". but clinicians love these diagnoses and believe they have utility despite not being scientific so there is no taste to get rid of them. Even the proposal to reduce the number of personality disorders in DSM-5 from 10 to 6 was ceremoniously rejected.