Security●●●Banger
Grantex–Open authorization protocol for AI agents(IETF draft submitted)
OAuth moment for agents—IETF draft submitted, standardizing what's currently API-key chaos.
Zero to OneBig BrainSolve My Problem
mishrasanjeev
203mo ago
Back to browse
ANML – A machine-first markup language for the agentic web (IETF Draft)
by ajeskey·May 21, 2026·7 points·4 comments
AI Analysis
●MidBold BetZero to One
ANML proposes a new protocol for agents, but adoption depends on an unproven agentic web.
Strengths
- •Dual XML and JSON serialization supports both human authoring and programmatic generation pipelines.
- •Built-in privacy constraints control field disclosure before agents share sensitive user data.
- •Explicit workflow state machines define multi-step interactions like booking flows deterministically.
Weaknesses
- •No reference implementation or live demo exists yet; currently just an IETF draft specification.
- •Solves a coordination problem that doesn't fully exist until autonomous agents broadly deploy.
Category
Target Audience
Protocol designers, AI agent developers, web standards contributors
Similar To
ActivityPub · OpenAPI Specification · W3C Web Things
Similar Projects
AI/ML○Pass
AEF – Agents State Machine
Two-commit spec repo with no working implementation or demo.
Ship It
mikemasam
103mo ago
Infrastructure●●Solid
Open Memory Specification (OMS), Context Assembly Language (Cal)
Content-addressed agent memory with append-only CAL—delete structural impossibility, not policy.
Big BrainNiche GemZero to One
sathishmg
103mo ago
AI/ML●Mid
Beam Protocol – SMTP for AI Agents (natural language agent-to-agent)
Agent addressing and discovery like email, but only 4 live agents and no adoption beyond creators.
Bold BetBig Brain
alfridus
323mo ago
Security●●Solid
Shadowscan – see what an AI agent can access on your machine
Scans MCP configs and env vars locally where other secret scanners miss agent context.
Solve My ProblemNiche Gem
sravyavedantham
113mo ago
AI/ML●●●Banger
Statewright – Visual state machines that make AI agents reliable
Protocol-level tool enforcement prevents jailbreaks better than system prompts.
Big BrainZero to OneSlick
azurewraith
301mo ago