A job-application tool should explain its evidence and its limits.
Use supplied evidence
Suggestions should reorganize, clarify, or strengthen facts the job seeker provided. Titles, dates, qualifications, employers, and results must not be invented.
Evaluate one role at a time
Keyword and evidence feedback is most useful when it is tied to one job description, company context, role, and country rather than a universal score.
State what a signal means
Scores and confidence labels are diagnostic estimates. They are not employer decisions, interview probabilities, or guarantees that an ATS will rank a resume.
Catalog scope is not presented as user success.
Seven boundaries govern the public product.
What the score examines—and what it cannot know
RoleSharp inspects extractable text, standard section structure, role and job-description term coverage, evidence quality, and selected market conventions. Real employers configure different systems and rules, so no score can reproduce every ATS or recruiter decision.
- Parsing and section readability
- Job-description term coverage
- Evidence and quantified outcomes
- Country-aware format guidance
Grounded edits remain traceable to the source
A useful suggestion should have a visible basis in the resume, profile, or job posting. If a requirement is not supported, RoleSharp should identify the gap rather than manufacture an answer.
- No invented experience or metrics
- One proposed change can be reviewed at a time
- Unsupported requirements stay unsupported
- Fictional demonstrations are labeled
Local conventions are guidance, not law
Country context can change expectations around page length, photos, personal details, and terminology. RoleSharp exposes that advice for review and override; job-posting instructions and applicable law take priority.
- Market context remains visible
- Guidance can be overridden
- High-risk personal details are not assumed
- Catalog coverage is separated from verified outcomes
Automation stops where evidence or consent stops
RoleSharp supports application-by-email only when a current posting explicitly publishes the recipient and instruction. It does not complete job-board or ATS forms, guess contacts, or call provider acceptance confirmed delivery.
- Posting-backed recipient required
- Consent, caps, cooldowns, and duplicate checks
- Immutable package remains reviewable
- Unknown outcomes cannot automatically retry
Mailbox permission and product login remain separate
Connected mail is limited to the documented send-only flow. RoleSharp does not request inbox-read, mailbox-modify, or contacts access for controlled email applications. Account deletion and retention boundaries are documented in the privacy policy.
- Separate delegated mailbox consent
- No message bodies or tokens in queue payloads
- Same-origin and encrypted-session controls
- Deletion controls remain accessible
Access terms and usage allowances answer different questions
Premium is purchased once for a stated number of days. During that term, tailored-resume generations use a rolling monthly allowance. The pass never auto-renews; when it expires, the account returns to Free while saved resumes remain available.
- 30-day and 365-day fixed terms
- One-time payment with no renewal charge
- Rolling generation allowance stated separately
- Saved work remains after expiry
Volatile comparisons are dated and fail closed
Pricing, free access, and product limits can change. Comparison evidence carries capture and review dates, volatility, source, and status. Expired, disputed, or non-equivalent evidence should not produce a numerical winner.
- One governed source for HTML and Markdown
- High-volatility claims require review dates
- Disputed evidence renders neutrally
- Corrections propagate across public formats
Specific corrections are reviewed against the source.
Send the affected URL, the exact statement, the supporting source, and why the current wording is incomplete or wrong. RoleSharp should correct the governed source first, then regenerate or update every derived surface.
Email a correction- High-intent errors take priority
- Disputed claims render neutrally during review
- Comparison corrections update HTML and Markdown
- Private resume or job text is not needed
Accountable contact
RoleSharp content and product decisions are accountable to Mustak, Founder of RoleSharp. Support and correction requests can be sent to support@rolesharp.com.
See the method applied to a fictional resume and job.
The homepage demonstration shows how source evidence, job requirements, grounded edits, diagnostic scoring, application controls, and tracking remain connected.