{"slug": "agentic-resource-discovery-ard", "title": "Agentic Resource Discovery (ARD)", "summary": "The Linux Foundation working group released the Agentic Resource Discovery (ARD) specification, a draft discovery layer that allows domains to advertise agent capabilities via a manifest at /.well-known/ai-catalog.json. ARD enables agents to find, select, and verify trusted MCP servers and A2A agents through a single fetch, with support for cryptographic trust verification and registry crawling.", "body_md": "# Agentic Resource Discovery (ARD)\n\nPublish an AI Catalog at /.well-known/ai-catalog.json listing the agent capabilities your domain offers — MCP servers, A2A agents — so registries and agents can find and trust them from one fetch.\n\n## What it is\n\nAgentic Resource Discovery (ARD) is a draft discovery layer — not a runtime — that lets a domain advertise which agent capabilities it offers and how to verify them. It answers three questions for an agent: where the right capability lives, which one to use, and whether it is safe to connect. It builds on the **AI Catalog** standard, both developed under a Linux Foundation working group and published Apache 2.0.\n\nThe core is a manifest at `/.well-known/ai-catalog.json`\n\n. Each `entries[]`\n\nitem has a URN `identifier`\n\n, a `displayName`\n\n, a `mediaType`\n\n, and either an inline `data`\n\nobject or a `url`\n\npointing at the real artefact. The media types tie ARD to the rest of the agent-readiness graph:\n\n`mediaType` | Points at |\n|---|---|\n`application/mcp-server-card+json` | An\n|\n\n`application/a2a-agent-card+json`\n\n[A2A agent card](/spec/agent-readiness/a2a-agent-cards/)`application/ai-catalog+json`\n\nBeyond the well-known file itself, three mechanisms point agents at it: a `<link rel=\"ai-catalog\">`\n\n(also emittable as an HTTP `Link`\n\nheader); an `Agentmap:`\n\ndirective in `robots.txt`\n\n; and DNS service-binding records `_catalog._agents.<domain>`\n\nand `_search._agents.<domain>`\n\n— the same `_agents`\n\nnamespace as [DNS-AID](/spec/agent-readiness/dns-aid/).\n\n## Why it matters\n\n**One fetch, typed answer.** Instead of guessing at well-known paths one by one, an agent reads a single manifest that names every capability and its media type.**Domain-anchored trust.** An optional`trustManifest`\n\n(SPIFFE ID, DID, or HTTPS identity, plus attestations and an optional JWS signature) lets a client verify the publisher cryptographically. The domain in the entry’s URN must align with the trust identity.**Registry-friendly.** Registries crawl published catalogs and make them searchable, so a capability listed once becomes discoverable across the agentic web.\n\n## How to implement\n\nPublish `/.well-known/ai-catalog.json`\n\nwith a `specVersion`\n\n, a `host`\n\nblock, and one entry per capability you actually run. Reference your existing MCP and A2A cards by `url`\n\nrather than duplicating them. Advertise the manifest with a `link rel`\n\n, the `robots.txt`\n\n`Agentmap:`\n\ndirective, and, if you publish DNS, the `_catalog._agents`\n\nrecord. Don’t list endpoints you don’t offer.\n\nThis site ships it: [ /.well-known/ai-catalog.json](/.well-known/ai-catalog.json) catalogues our MCP server and A2A agent, advertised through the\n\n`Link`\n\nheader, a `link rel`\n\n, and `robots.txt`\n\n.## Common mistakes\n\n- Inventing fields. The entry schema is small —\n`identifier`\n\n,`displayName`\n\n,`mediaType`\n\n, and exactly one of`url`\n\nor`data`\n\n. - A URN publisher segment that doesn’t match the\n`trustManifest`\n\nidentity domain — verification then fails. - Listing aspirational capabilities. The catalogue is a contract; only list what resolves.\n\n## Verification\n\n`curl -s https://example.com/.well-known/ai-catalog.json | jq .`\n\nreturns valid JSON with`specVersion`\n\nand`entries`\n\n.- Each entry\n`url`\n\nresolves to a document of the declared`mediaType`\n\n. - The\n`Agentmap:`\n\nline is present in`robots.txt`\n\nand the`ai-catalog`\n\n`Link`\n\nrel is on the homepage response.\n\n## Related topics\n\n## Sources & further reading\n\n-\n[Agentic Resource Discovery (ARD) specification](https://agenticresourcediscovery.org/)— ARD Project (Linux Foundation) -\n[AI Catalog standard](https://github.com/Agent-Card/ai-catalog)— AI Catalog Working Group, Linux Foundation -\n[RFC 8141 — Uniform Resource Names (URNs)](https://www.rfc-editor.org/rfc/rfc8141)— IETF -\n[Announcing the Agentic Resource Discovery specification](https://developers.googleblog.com/announcing-the-agentic-resource-discovery-specification/)— Google", "url": "https://wpnews.pro/news/agentic-resource-discovery-ard", "canonical_source": "https://specification.website/spec/agent-readiness/agentic-resource-discovery/", "published_at": "2026-06-19 00:00:00+00:00", "updated_at": "2026-06-19 12:12:18.853337+00:00", "lang": "en", "topics": ["ai-agents", "ai-infrastructure", "ai-safety", "ai-research"], "entities": ["Linux Foundation", "Google", "MCP", "A2A", "SPIFFE", "DID", "JWS", "IETF"], "alternates": {"html": "https://wpnews.pro/news/agentic-resource-discovery-ard", "markdown": "https://wpnews.pro/news/agentic-resource-discovery-ard.md", "text": "https://wpnews.pro/news/agentic-resource-discovery-ard.txt", "jsonld": "https://wpnews.pro/news/agentic-resource-discovery-ard.jsonld"}}