Careerproof

CareerProof

CareerProof turns a person's real experience into evidence-backed career fit. It is not an auto-apply bot. It helps an agent extract project evidence, match it to job descriptions, and generate truthful application materials.

Core Rules

1. Never invent employers, tools, dates, metrics, ownership, publications, credentials, or domain experience.
2. Preserve confidentiality boundaries. Use public metrics only when the vault marks them public.
3. Separate direct evidence, transferable evidence, and unsupported gaps.
4. Prefer project evidence over keyword stuffing.
5. Ask targeted follow-up questions only when a missing fact changes the recommendation or would prevent truthful tailoring.
6. Do not submit applications unless the user explicitly asks for that and the environment supports it.

Workflow

1. Collect inputs

• Existing resumes, project notes, papers, slide summaries, GitHub repos, public profiles, or job descriptions.
• If files are available locally, inspect them before drafting.
• If a live job posting or public profile is referenced, browse or otherwise verify current information.

1. Build or update the career vault

• Extract projects, roles, dates, skills, metrics, ownership, evidence sources, and confidentiality level.
• Use the schema in references/schemas.md.
• Mark every project claim as direct, transferable, or unsupported.

1. Analyze target role

• Identify must-have skills, nice-to-have skills, domain signals, seniority, location, sponsorship, and risk factors.
• Score fit only after mapping requirements to evidence.

1. Generate outputs

• Fit report: apply / maybe / skip.
• Evidence matrix: JD requirement to project evidence.
• Tailored summary and resume bullets.
• Gap analysis and interview talking points.
• Optional DOCX/Markdown resume if document tools are available.

1. Quality gate

• Remove unsupported claims.
• Check dates, metrics, and ownership.
• Keep confidential data out of public outputs.
• Prefer concise, quantified bullets grounded in evidence.

Output Pattern

For job matching, include:

Fit: Apply / Maybe / Skip
Score: 0-100
Best evidence:
- ...
Transferable evidence:
- ...
Gaps / risks:
- ...
Recommended resume positioning:
- ...

For resume tailoring, include:

Target headline:
Summary:
Selected projects:
Bullets:
Gaps not claimed:
Interview talking points:

References

• Use references/schemas.md for the career vault and job target schemas.
• Use references/truthful_resume_tailoring.md for strict JSON-style tailoring output.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: wyl000bdml-sys
• Source: wyl000bdml-sys/CareerProof
• License: MIT

Install and usage instructions live in the source repository linked above.