FAQ: Harvard fit, scores, and automated checks

Harvard fit checklist

On the builder, the sidebar list titled Harvard fit checklist compares your draft to common Harvard-style résumé conventions. Each row shows a green check when the rule passes, or an amber alert icon when it fails or needs your judgment. A second line (smaller, gray text) gives a metric or hint, for example what percentage of bullets include a digit.

Clear Harvard section order

Passes when section order is the default: Education, Experience, Leadership & activities, Skills & interests. If you reorder sections, the checklist reminds you to confirm the order still matches what you want.

Reverse chronology (manual review)

Always marked for manual review. The app does not sort your jobs or degrees; you should put the newest items first in each list.

Quantified bullets where possible

Looks at every non-empty bullet in Experience and Leadership. A line "counts" as quantified if it contains any digit (0–9), for example a year, a percentage, or a count. Passes when at least 35% of bullets include a digit, or when there are no bullets yet.

Concise wording

Fails if any bullet is longer than 180 characters, to help keep one-page density.

Professional summary length

If the summary is empty, it passes (summary is optional). If you fill it, pass is when length is between 50 and 700 characters.

Contact header formatting

Passes when your name has more than one character and at least two of: email, phone, city/region are filled. Email is also checked for a simple valid shape when present.

Strong action-verb openings

Each bullet's first word is checked against a list of strong verb stems (e.g. Led, Built, Designed) and a small heuristic (past-tense -ed verbs). Passes when at least 50% of bullets match, or when there are no bullets.

Avoid first-person pronouns

Fails if any bullet contains common first-person words (I, me, my, we, our, etc.).

Avoid weak opening words

Fails if a bullet's first word (after stripping punctuation) is one of: responsible, helped, worked, handled, assisted, duties, various.

Review panel

Above the checklist, Review lists practical warnings: too many bullets in one role (more than five), bullet length and pronouns again, missing education or experience before export, and a rough size hint if the draft might overflow one printed page.

Experience: percentage scores

Next to each Experience field label you may see a 0–100% score. Color is a quick signal: green (about 85%+), amber (about 55 to 84%), or red (below ~55%). Scores favor completeness (dates in YYYY-MM form, employer and title length) and, for tools, having more than one comma-separated item when you choose to list tools.

Under many fields you will also see the same checklist-style row as in the sidebar (icon + title + gray detail). That row explains the main pass/fail reason for that field or bullet group; it is the same idea as the sidebar, not a second unrelated system.

Per-bullet checks

Each achievement line is scored with the same building blocks as the checklist: action-verb opening, weak first word, first person, length over 180 characters, and a simple "long line without any digit" hint for quantification. The percentage on the bullet is a composite of those deductions, capped at 0–100.

The Bulletsheader shows an average of those scores across non-empty lines, plus a section checklist row that mirrors "Quantified bullets" and "Weak opening words" using counts from your bullets.

JSON import

Optional Show JSON import exposes a prompt you can copy into any tool that outputs JSON, plus a paste box for the array. Parsing strips common markdown fences and can find an array inside surrounding text. Imported roles append to Experience; scoring then runs on the new text like any other edit.

Save JSON draft at the top of the builder stores a full snapshot in local storage, separate from scores, for backup or moving between machines.