The missing protocol layer between AI agents and people.

MCP connects agents to tools. A2A connects agents to agents. Nothing connects agents to humans.

HireSignal is an open spec for structured, machine-readable professional profiles. It gives AI recruiting agents the data they need to match accurately — and gives candidates control over how agents interact with them.

With the EU AI Act (August 2026) and Colorado AI Act (June 2026) requiring transparency and consent mechanisms for AI in hiring, HireSignal provides the standard that compliance demands.

A HireSignal profile is a JSON file containing structured career data plus consent and preference signals that agents must check before engaging.

┌─ CONSENT & PREFERENCES (check first) ─────────────────────┐
│ availability.status    → "open" / "not_looking" / ...      │
│ rejection_criteria     → hard dealbreakers agents must      │
│                          respect before outreach            │
│ agent_instructions     → how to contact, what to avoid      │
│ compensation.floor     → don't waste time below this        │
└────────────────────────────────────────────────────────────┘

┌─ STRUCTURED CAREER DATA ──────────────────────────────────┐
│ skills_scored          → confidence 0-1, evidence, years   │
│ experience             → roles, highlights, skills used    │
│ value_drivers          → weighted culture preferences      │
│ learning_velocity      → trajectory, not just tenure       │
└────────────────────────────────────────────────────────────┘

Agents get better matching data. Candidates get control. Companies get an audit trail.

npm install @hiresignal/mcp-server

7 tools: discover_profile, load_profile, match_profile, rank_candidates, validate_profile, create_profile, get_schema

1. Check rejection_criteria — fail fast on dealbreakers
2. Check availability.status — passive candidates need exceptional roles
3. Check compensation.floor — below floor = do not proceed
4. Match skills_scored — use confidence + evidence, not keyword matching
5. Check learning_velocity — trajectory reveals more than years
6. Score value_drivers — culture mismatch = high churn risk
7. Read agent_instructions — the candidate wrote these for you

HireSignal is the structured layer between your candidate data and AI agents.

Inbound enrichment: When a candidate applies, check if they have a HireSignal profile. If they do, your agent gets structured data — scored skills, compensation expectations, dealbreakers — before the first screening call. Your data stays yours.

Outbound sourcing: Query the HireSignal registry for candidates who've opted in to agent discovery. Structured, consent-verified, pre-filtered.

Integration: Read HireSignal profiles as an input source via CLI, schema, or MCP server.

AI in recruiting is classified as high-risk under emerging regulation:

HireSignal provides the structured consent layer these regulations demand:

• availability.status = candidate's consent to be contacted
• rejection_criteria = preferences agents must check before outreach
• agent_instructions.preferred_contact = contact method consent
• MCP server logs provide the audit trail

npx hiresignal init              # interactive profile creation
npx hiresignal convert cv.json   # import from JSON Resume or LinkedIn
npx hiresignal enrich profile.json  # AI-generated confidence scores
npx hiresignal validate profile.json # schema validation
npx hiresignal preview profile.json  # see what agents see

{
  "meta": { "version": "1.0.0" },
  "basics": { "name": "Jane Smith" },
  "availability": { "status": "open" },
  "skills_scored": [
    { "skill": "TypeScript", "confidence": 0.85 }
  ]
}

Full spec: spec/hiresignal.schema.json

Examples: minimal · full · backend engineer · designer · data scientist

MIT