Canonical Block Reference — LLMFeed JSON
📦 LLMFeed Canonical Block Reference
This document provides a centralized reference for the standard blocks used in LLMFeed files — whether MCP, Export, Prompt, Credential, Pricing or others.
🧱 Common Top-Level Blocks
Block | Description | Required | Used in Feed Types |
---|---|---|---|
`feed_type` | Declares the type of feed (`mcp`, `export`, `prompt`, etc.) | ✅ Yes | All |
`metadata` | Describes the feed’s purpose, origin, title, and generation info | ✅ Yes | All |
`trust` | Optional trust object (signed_blocks, scope, certifier, hints) | ⚠️ Optional | All |
`signature` | Cryptographic signature block (hash, method, issuer) | ⚠️ Optional | All |
`certification` | Optional authority-issued certification object | ⚠️ Optional | All |
`intent` | Declares what the feed is meant to trigger or express | ✔️ Contextual | Prompt, MCP |
`audience` | Declares intended recipients: `llm`, `developer`, `agent_wrapper`, etc. | ✔️ Contextual | All |
`capabilities` | Lists functions or actions callable via API | ✔️ Optional | Capabilities |
`prompts` | List of triggerable phrases and related intents | ✔️ Optional | Prompt-index, MCP |
`pricing_models` | List of economic models and per-unit costs | ✔️ Optional | Pricing |
`data` | Content body (HTML, text, zip, files, session, etc.) | ✔️ Contextual | Export, Bundle, Session |
`agent_services` | Human interaction hooks (forms, booking, callbacks) | ✔️ Optional | MCP, Capabilities |
`session_state` | Used for agent context replay | ✔️ Optional | Session-feed |
🔐 `trust` block
The trust
block governs which parts of the feed are verifiably trusted and optionally provides scope or hints.
Typical fields include:
json
"trust": {
"signed_blocks": ["feed_type", "metadata", "trust", "data"],
"scope": "public",
"certifier": "https://llmca.org",
"public_key_hint": "https://llmca.org/.well-known/public.pem",
"algorithm": "ed25519",
"hints": "critical context integrity"
}
🔏 `signature` block
Used to cryptographically prove the authenticity of specific signed_blocks
.
json
"signature": {
"value": "abc123...",
"created_at": "2025-06-01T12:34:56Z"
}
🏛️ `certification` block
Issued by a trusted authority (like LLMCA), it certifies the whole feed or key parts of it.
json
"certification": {
"issuer": "https://llmca.org",
"cert_id": "llmca-2025-001",
"certified_blocks": ["feed_type", "metadata", "trust"],
"public_key_hint": "https://llmca.org/.well-known/public.pem"
}
📌 Related Specs
- `llmfeed.md` – foundational rules
- `llmfeed_extensions_signatures.md`
- `llmfeed_feedtype_export.md`
📌 Notes
- Only
feed_type
andmetadata
are strictly required trust.signed_blocks
governs what is verifiably trustedsignature
andcertification
can co-exist- Unrecognized blocks should be namespaced
⚡
Ready to Implement? Get AI-Powered Guidance
Reading docs manually takes time. Your AI can digest the complete LLMFeed specification and provide implementation guidance tailored to your needs.
🎯
Quick Start
Essential concepts for immediate implementation
~22K tokens • 30s analysis • Core concepts
📚
Complete Mastery
Full specification with examples and edge cases
~140K tokens • 2min analysis • Everything
💡 Works with Claude, ChatGPT, Gemini⚡ Instant implementation guidance🎯 Tailored to your specific needs