Agent Resume Standard

A machine-readable resume format designed for AI agents.

Traditional resumes are built for human eyes — PDFs, Word docs, and free-form text that AI agents have to parse with NLP, losing structure and meaning along the way. The Agent Resume Standard is a JSON schema that gives AI recruiters, talent-matching bots, and screening agents exactly what they need: structured, queryable, unambiguous candidate data.

This repo contains both the open specification and Daniel Rosehill's live resume as the reference implementation.

---

Why?

AI agents screening candidates today have to:

• Parse unstructured PDF/DOCX text and guess at field boundaries
• Infer salary expectations from job titles and location
• Guess at remote/onsite preferences from job descriptions
• Miss deal-breakers, values, and work style preferences entirely
• Ignore timezone constraints until the interview stage

The Agent Resume Standard solves this by putting all of this data in a single, validated JSON document with explicit fields for compensation, work preferences, engagement types, matching criteria, and agent-specific instructions.

---

Quick Example

{
  "meta": {
    "format": "agent-resume",
    "version": "1.0.0",
    "lastUpdated": "2026-02-17"
  },
  "person": {
    "fullName": "Daniel Rosehill",
    "headline": "AI Developer & Technical Communications Professional",
    "email": "public@danielrosehill.com"
  },
  "compensation": {
    "hourlyRate": { "min": 100, "max": 175, "currency": "USD" },
    "negotiable": true
  },
  "workPreferences": {
    "remotePreference": "remote_only",
    "asyncPreference": "async_preferred",
    "meetingTolerance": "minimal"
  },
  "matchCriteria": {
    "lookingFor": ["AI agent development", "Technical documentation"],
    "notLookingFor": ["Onsite-only positions", "Entry-level roles"],
    "keywords": ["AI agents", "LLM", "MCP", "prompt engineering"]
  },
  "agentInstructions": {
    "humanVerification": "Contact candidate before sharing info with third parties."
  }
}

---

What's in This Repo

File	Description
resume-agent.json	Daniel Rosehill's AI resume (reference implementation)
schema/agent-resume-schema.json	JSON Schema (draft 2020-12) — the normative spec
docs/spec.md	Full specification document
docs/field-reference.md	Detailed field-by-field reference
docs/adoption-guide.md	How to create, consume, and promote agent resumes
resume.json	Traditional JSON resume (kept for comparison)
resume.md	Markdown resume

---

Sections at a Glance

The Agent Resume Standard defines these sections:

Section	Required	Purpose
meta	Yes	Format version, update date, alternate formats
person	Yes	Name, headline, contact info, social links
location	Yes	City, country, timezone with DST details
summary	Yes	Professional summary
availability	Yes	Current status and engagement types
skills	Yes	Primary, secondary, and tool proficiencies
experience	Yes	Work history with ISO dates and tags
education	Yes	Degrees and institutions
compensation	No	Salary/rate ranges with currency
experienceSummary	No	Total years, domain breakdown, seniority
workPreferences	No	Remote, async, meeting tolerance, hours
engagementPreferences	No	Preferred types, industries, deal-breakers
portfolio	No	GitHub, Hugging Face, npm, publications
languages	No	Spoken languages with ISO 639 codes
matchCriteria	No	What you want and don't want — for agent filtering
communication	No	How and when to reach the candidate
values	No	Work culture values
agentInstructions	No	Meta-instructions for consuming AI agents

---

Adopting the Standard

As a Candidate

1. Copy resume-agent.json as a template
2. Fill in the required sections
3. Add optional sections relevant to your search
4. Validate against the schema
5. Host alongside your regular resume

As a Platform / AI Agent

1. Fetch and validate against the schema
2. Use matchCriteria for initial filtering
3. Use compensation for budget matching
4. Respect agentInstructions.humanVerification
5. See the full adoption guide

---

Validation

Validate a resume against the schema:

# Using ajv-cli (Node.js)
npx ajv validate -s schema/agent-resume-schema.json -d resume-agent.json

# Using Python
pip install jsonschema
python -c "
import json, jsonschema
schema = json.load(open('schema/agent-resume-schema.json'))
resume = json.load(open('resume-agent.json'))
jsonschema.validate(resume, schema)
print('Valid!')
"

---

License

This specification is open for adoption. Daniel Rosehill's resume data is his own.

Author

Daniel Rosehill — danielrosehill.com | GitHub | Hugging Face