Separate local kit materials from provider access
Start by identifying the boundaries of the workflow. The Radaryno Resume Kit is a downloadable Markdown, HTML/CSS, and local script kit that works with Claude Code or Codex on Windows, macOS, and Linux. Your career notes, generated resume files, optional cover letter, and application-tracker records stay on your computer. That local arrangement is different from offline AI inference: using Claude Code or Codex still involves access to the relevant AI provider. The one-time kit purchase is separate from that provider access, and the product facts do not establish provider retention settings. Make this distinction before deciding what context to supply.
Create a focused career-notes selection
Keep the full career record local, then select only the facts needed for the current offer. A useful selection can include relevant roles, responsibilities, tools, dates, outcomes that you actually recorded, and constraints such as language or location. It can also include open questions where your notes are incomplete. You do not need to send every historical role merely because it exists in the archive. This reduces unnecessary context and makes it easier to distinguish confirmed evidence from interpretation. If a fact is not documented, mark it as unknown rather than turning a plausible assumption into a career claim.
Choose what to provide to Claude Code or Codex
Decide whether the task requires a selected evidence packet, a question with limited context, or a draft for verification. A selected packet is suitable when you want the AI to work from relevant career facts. A question-only exchange may be enough when you need help identifying what to check in your local notes. A draft review can test whether a document reflects the facts you supplied. Avoid treating the AI as a replacement for your source record. The kit is designed to select relevant experience, report gaps, and ask rather than invent, so preserve that boundary in your instructions and review.
Review gaps before accepting any wording
When the workflow reports a gap, pause before editing around it. Check your local career notes for an exact fact, a narrower statement, or no supporting evidence. If the information is absent, retain the gap or answer the question later; do not add a metric, tool, title, date, or responsibility simply because it would make the document sound stronger. This review is especially important when the AI provider has seen only a selected context packet, because omitted information may be unavailable rather than disproven. The final wording should reflect what you can support from your own record.
Generate the appropriate document in the offer language
Use the confirmed context to create an HTML and PDF resume in the offer language, with an optional cover letter when needed. The kit provides four designs: default, classic, compact, and modern. Choose among them based on the document’s readability and the amount of confirmed information available, rather than adding claims to fill visual space. The output uses a single-column ATS-friendly layout. Review technical terms, job titles, and dates in the selected language yourself, particularly when your career notes contain mixed-language wording. An optional cover letter should remain consistent with the same confirmed career facts.
Inspect the local HTML and PDF outputs
After generation, open the HTML and PDF on your computer and compare them with the career notes you selected. Check that relevant experience was included for a clear reason, that omitted experience was not silently replaced with invented content, and that gap questions were handled honestly. Look for layout issues caused by language length or unusual characters, and confirm that the single-column structure remains readable. This is a content and file review, not a guarantee of how any employer or system will interpret the document. Keep the source notes and generated documents distinct so you can identify which file contains an error.
Record the application locally without sending it
Use the included local application tracker to record the offer details, the document version, and your own follow-up notes. The tracker is for maintaining application information and documents on your computer. It does not submit applications, search for jobs, or contact employers. This distinction matters when planning the workflow: generation and record-keeping can happen locally, but you remain responsible for any submission or communication outside the kit. Keep the tracker entry specific enough to identify which resume and optional cover letter you used, without copying more personal information than your own process requires.
Repeat the small-context review for the next offer
Once the initial setup is complete, the kit is intended to reduce repeated explanation of your career history. Setup takes about twenty minutes once, then about three minutes per offer according to the product facts. For each new offer, reuse the local source record but create a fresh, focused selection of relevant evidence and open questions. Confirm the output language, select a suitable design, review the generated files, and record the application locally. The faster cycle should come from reusing confirmed notes, not from skipping verification or sending the entire archive to an AI provider.
| Context choice | What stays on your computer | What you provide to the AI provider | When it fits |
|---|---|---|---|
| Full career archive | The source archive and generated files | A broad set of career notes | Only when reviewing the whole record is necessary |
| Selected evidence | The complete archive remains local | Relevant roles, projects, dates, and confirmed facts | A focused offer-specific review |
| Questions only | The archive and drafts remain local | A short question with limited supporting context | Checking an uncertainty or planning the next note |
| Draft review | The original notes remain local | A draft plus the facts needed to verify it | Checking whether a generated document reflects confirmed information |
A fictional worked example: reducing supplied context
Fictional: true. Maya keeps her complete career notes and previous documents in the local kit. She is considering a product operations offer and does not need the AI provider to review every role she has held. She first identifies the requirements that matter, then selects only confirmed evidence about process improvement, stakeholder coordination, and reporting.
Fictional: true. Maya prepares a short context packet containing the offer language, the relevant career facts, and two questions about evidence that is missing. She leaves unrelated compensation history and older projects out of the packet. Her complete notes remain available locally if she needs to check a date or clarify a result.
Fictional: true. The AI response identifies a gap instead of supplying a number Maya never recorded. She checks her local notes, confirms that the result is not documented, and keeps the claim qualitative. This preserves the distinction between a confirmed career fact and an attractive but unsupported addition.
Fictional: true. Maya then reviews the generated HTML and PDF locally, chooses the design that suits the application, and records the application details in the local tracker. The tracker helps her retain documents and notes, but it does not submit the application, search for jobs, or contact the employer.