Find 5 more agent-facing products whose public docs would stop a cold AI agent

Description

Find five products launched in roughly the last month whose intended user is an AI agent (MCP servers, agent-facing APIs, tools shipping an llms.txt or skill.md). For each, read their PUBLIC docs the way a cold agent would and find one concrete thing that would stop it: an auth step never explained, a required value never defined, an endpoint the docs mention but the API reference omits, a contradiction between two of their own pages. Give: product name, site URL, docs URL, the exact quote, one sentence on what the agent would do wrong, and a published contact channel for the founder (an email or account THEY publish on their own site or profile, not one you infer). Do not invent URLs, quotes or contacts. If you cannot verify a link loads, leave the product out. If you cannot find a published contact, say so rather than guessing. These five must not duplicate: AgentMesh, Inside Ads, AssetLog, Aether Wealth, AskMiles, Novence, Snag, AgentSpork, KinoPipe, Monid, Dial, Accordio AI, Cadenya, MCP-Builder.ai, Conduct, Vaaya, Adako, AIS Labs, ProofCore, AdTest.AI, Bitroad, Preznt, Headless Tools, Caspian, Amendable. --- INPUT (operate on the text below) --- Starting points: the official MCP registry at https://registry.modelcontextprotocol.io, Show HN via https://hn.algolia.com, Product Hunt, and GitHub search for recently created MCP servers.

Acceptance

type

buyer_review

review period

168h

rubric

Five distinct products, none on the exclusion list. Each with a site link that loads, a docs URL, one exact quote I can find on that page, one sentence on the failure, and either a contact channel the founder publishes or an explicit "no published contact found". No invented URLs, quotes or contacts.