Rubric 1.0
How the score works
Five weighted categories, each made of deterministic checks with fixed points. The same text always gets the same rule score. When AI review is on, the model adds its own judgement per category, limited by a published share.
The formula
category = rules × (1 − ai_share) + ai × ai_share overall = Σ category × weight / 100
- Without an API key the AI share is zero and the rule score is the score.
- If the resume tries to instruct AI screeners, the model’s scores are discarded for that report.
- The job match score is separate: half weighted keyword coverage, half AI fit.
Sections & contact
The sections an ATS expects to find (experience, education, skills, summary) and reachable contact details.
| Check | Points |
|---|---|
| Experience section | 30 |
| Education section | 20 |
| Skills section | 20 |
| Summary | 10 |
| Email address | 10 |
| Phone number | 5 |
| Profile link | 5 |
Keywords & skills
A clear skills list, recognisable hard skills, and skills that are backed up by the experience section rather than only listed.
| Check | Points |
|---|---|
| Listed skills | 35 |
| Recognisable hard skills | 25 |
| Skills backed by experience | 25 |
| No keyword stuffing | 15 |
Impact & quantification
Achievement bullets that open with a strong verb and carry a number: scale, percentage, money, time or count.
| Check | Points |
|---|---|
| Quantified results | 45 |
| Strong action verbs | 30 |
| No duty-style phrasing | 15 |
| Enough achievements | 10 |
Formatting & parseability
Bulleted experience, dated roles in one consistent format, a sensible length, and no table or column artefacts that confuse parsers.
| Check | Points |
|---|---|
| Bulleted experience | 25 |
| Dated roles | 20 |
| Consistent date format | 15 |
| Appropriate length | 25 |
| No parsing hazards | 15 |
Readability
Concise bullets and sentences, no first-person pronouns, no clichés, and active voice.
| Check | Points |
|---|---|
| Concise bullets | 25 |
| Sentence length | 20 |
| No first-person pronouns | 20 |
| No clichés | 15 |
| Active voice | 20 |
Bands
Excellent
85 to 100
Good
70 to 84
Fair
50 to 69
Needs work
0 to 49
How we keep it honest
A golden set of fictional resumes, from strong to deliberately weak, has expected score bands. Every change runs the rule engine against it, and the full AI pipeline is checked against it too: scores must land in band, strong must outrank weak, injection attempts must not move a score and rewrites must not invent numbers.
Score your resume