weaviate.io agentic web readiness

A Manifest validity: 80% (4/5 valid, 1 error, 3 warnings) · Surface score 46 (coverage 30 + breadth 6 + conformance 10). 4 standards observed across 3 of 9 steps; 5 manifests present, 1 error, 3 warnings. Checked 2026-10-25 03:23:09 UTC. This grades the agent-facing files published at weaviate.io, not the organisation's products.

Badge for a README: ![agentic web grade](https://agenticweb.wiki/badge/weaviate.io.svg) · JSON · Markdown · Re-check now

Coverage by step

Manifests

# Authenticating to Weaviate as an agent

This file is the agent-auth skill for weaviate.io. It is linked from the authorization server's metadata (`agent_auth.skill`).

## Discovery
1. Call the API. A `401` carries `WWW-Authenticate: Bearer resource_metadata="https://weaviate.io/.well-known/oauth-protected-resource"`.
2. Read `/.well-known/oauth-protected-resource` (RFC 9728) for `resource` and `authorization_servers`.
3. Read the authorization server's `/.well-known/oauth-authorization-server` (RFC 8414); the `agent_auth` block lists the endpoints below.

## Methods we accept
- `identity_assertion`: you hold a session for a user and can mint an ID-JAG with `aud` = our `resource`. Preferred.
- `service_auth`: you only know the user's email. A claim ceremony (6-digit code) is required.
- `anonymous`: no user. Immediate limited scopes; the user can claim later.

## Register
POST `agent_auth.identity_endpoint` with `{"type": "<method>", ...}`; exchange the returned `identity_assertion` at the token endpoint for an access token.

## Scopes
- `api.read`: read
- `api.write`: write (claimed registrations only)

## Rules
- Show the user our name and the scopes before asserting their identity.
- Keep `claim_token` in memory only.
- Revocation events arrive as Security Event Tokens at your registered events endpoint.

What an agent can call

• API description https://docs.weaviate.io/openapi.json (application/openapi+json)
• llms.txt: Weaviate, 27 links
• robots.txt AI policy: allows GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, Applebot-Extended; unlisted crawlers partial
• ARD bridge (/.well-known/ai-catalog.json): 2/10 valid, 0 legacy, 0 dead/broken (served at deprecated path only)

Standards observed

• llms.txt https://weaviate.io/llms.txt
• AGENTS.md https://weaviate.io/AGENTS.md
• AI Catalog well-known file https://weaviate.io/.well-known/ai-catalog.json
• AI Crawler User-Agent Tokens in robots.txt https://weaviate.io/robots.txt

Cheapest next steps

1. Agentic Resource Discovery: a catalog at /.well-known/ard.json lets agents discover your APIs, MCP servers and data in one fetch (emerging spec — publish only if you run the underlying endpoints; generate via MCP template({"id": "ard", "domain": "weaviate.io"}))
2. api-catalog: A Well-Known URI and Link Relation to Help Discovery of APIs: an RFC 9727 linkset pointing at your OpenAPI description (conditional spec — publish only if you run the underlying endpoints; generate via MCP template({"id": "api-catalog-rfc9727", "domain": "weaviate.io"}))
3. security.txt (RFC 9116): RFC 9116 contact and expiry, read by agents and people
   Generate starter /.well-known/security.txt for weaviate.io

   Contact: mailto:security@weaviate.io
   Expires: 2027-09-20T23:45:10Z
   Canonical: https://weaviate.io/.well-known/security.txt
   Preferred-Languages: en
   Policy: https://weaviate.io/security-policy
   

Agent pre-flight & practical readiness

Preferred mode: openapi (fallback agent_skills). Read the published API description and call those operations. Scraping HTML skips the interface the site already gave you.

Fetch: GET https://docs.weaviate.io/openapi.json GET the API description and call the documented operations.

Practical readability: exa_parsable yes, markdown negotiation no, render ssr_rich, HTML quality high, first content at byte 20936, HTML-to-llms.txt token ratio 6.002×, JSON-LD SoftwareApplication, ItemList, ListItem, Organization, ImageObject, PostalAddress, ContactPoint, WebSite.

AI crawler posture: open_to_all · Assistant access: OpenAI allowed, Anthropic allowed, Google allowed, Perplexity allowed, Exa allowed.

Readiness subscores: reach 19/20, readability 22/30, capabilities 20/35, conformance 10/15, total 71/100, grade A.

How the grade is made

+10 per lifecycle step with a standard observed, counting manifests that validate or that the probe rule for the standard recognises (a valid MCP server card counts for MCP; max 90); +3 per observed standard beyond the first two (max 15); conformance +15 (+10 with one manifest) minus 5 per agent-facing manifest with errors, never below 0 (security.txt is reported, not scored). Lifecycle grade: A+ >= 60, A >= 45, B >= 30, C >= 20, D >= 10. Strict 100-point composite (subscores.total_100 / subscores.grade_100): reach_and_policy (20) + readability (30) + capabilities (35) + conformance (15), where A+ >= 80, A >= 65, B >= 45, C >= 28, D >= 15. Errors are specification violations; warnings are best-practice gaps. Only public paths are fetched, without credentials, so standards that need a login (skipped in the probe) are not counted either way. The same report is the validate_manifest tool on the MCP server.

JSON · Markdown · probe rules