Tailor an existing resume to a specific role without inventing experience, flattening the candidate's voice, or silently modifying the source document.
Most resume generators start from a job description and produce plausible copy. That is also how unsupported metrics, inflated ownership, and generic keyword soup get introduced. Generate Resume starts from a different constraint: every material claim must trace to candidate-approved evidence.
Ask your agent:
Install the generate-resume skill from https://github.com/mifisher/generate-resume
Or install it with the Skills CLI:
npx skills add mifisher/generate-resume --skill generate-resume --global --yesProvide a job description, your baseline resume, and an evidence source such as an accomplishment bank or verified career history:
Use $generate-resume to tailor my baseline resume to this product lead role.
Job description: <URL or pasted text>
Baseline: <file or document link>
Evidence: <file, folder, or document link>
Constraints: one page, preserve the existing layout
The skill will:
- Extract the role's highest-value requirements.
- Map them to direct evidence, adjacent experience, or genuine gaps.
- Propose a reviewable change manifest with evidence for each material claim.
- Wait for approval before creating or changing a document.
- Fork the baseline, apply the approved edits, and verify the rendered result when tools permit.
It never submits an application.
- Evidence before polish. New metrics, credentials, and ownership claims require a traceable source.
- Gaps stay gaps. Transferable experience is framed honestly instead of being upgraded into direct experience.
- The baseline stays clean. The skill creates a fork and never edits the canonical resume in place.
- Approval before mutation. You see the summary, skills, bullet changes, reasons, and evidence before a final document is created.
- Rendered-output verification. Page count and formatting are claims to verify, not assumptions.
- Tool-neutral operation. Google Docs, Word, local files, and other document systems can work when the agent has the corresponding capability.
Job Search OS manages the broader workflow: role discovery, company records, pipeline state, follow-ups, and a resume edit manifest. Generate Resume handles the evidence-backed document transformation.
Job Search OS Generate Resume
discover or add a role read the role + evidence
↓ ↓
create company record map requirements to claims
↓ ↓
generate resume manifest ─────────▶ propose and approve edits
↓
record resume_ref ◀───────── create and verify the fork
They are independent projects. Use Generate Resume by itself with any baseline and evidence source, or connect it to a private Job Search OS instance using the included integration reference.
skills/generate-resume/SKILL.md— the agent workflow.evidence-and-claims.md— claim traceability rules.resume-rules.md— default editorial and ATS guidance.job-search-os-integration.md— optional handoff contract.
Resume data is sensitive. Keep real resumes, evidence banks, job-search records, recruiter information, and generated variants in private storage. This public repository contains no candidate data and should not be used as a storage location for application materials.
MIT