The Problem:

Agents evolve. The protocol has no way to communicate version relationships, deprecation timelines, or migration paths at discovery time.

Current State:

• Optional version field (string, no semantics)
• Optional updatedAt timestamp
• No deprecation indicators
• No "this replaces that" relationships

Scenario:

Invoice Processor v1.0 → v2.0 (breaking change)

1. The agent discovers v1.0 (or maybe it was cached)
2. Metadata indicates v1.0 EOL date: 2027-01-01
3. Metadata also proposes a replacement
4. Agent evaluates the replacement and decides whether to use it or keep searching an alternative (based on EOL data, it might use the deprecated one anyway if the workflow is short lived)

Proposal:

• Add an optional lifecycle object that tracks signficant SDLC events, including deprecation.
• The deprecation object not only provides dates, it also provides replacement recommendation and migrationGuide

{
  "identifier": "urn:ai:acme.com:finance:invoice-processor",
  "version": "2.0.0",
  "lifecycle": {
    "status": "stable",
    "releaseDate": "2026-03-15",
    "deprecated": {
        "replacedBy": "urn:ai:acme.com:finance:invoice-processor",
        "deprecationDate": "2026-09-01",
        "endOfLifeDate": "2027-01-01",
        "breakingChanges": [
          "Authentication changed from API key to OAuth2",
          "Response schema now uses ISO 8601 dates"
        ],
        "migrationGuide": "https://docs.acme.com/invoice-v2-migration"
       }
  }
}

Discovery Value:

• Coordinated rollouts: discover what's coming before it goes live
• Filter out deprecated agents at discovery time
• Understand breaking changes before selecting an agent
• Find migration guides during discovery phase