tp tate@programs pulse live
tate@programs ~/notes/mcp-registry-pulse 2026-06-24

public registry snapshot

MCP Registry Pulse.

A launch-readiness snapshot across current MCP Registry metadata and linked GitHub repos. This is aggregate analysis only: it shows where server launches tend to lose trust before a user ever connects a client.

generated
2026-06-24
registry rows
300
latest servers
129

topline

The launch bar is now metadata plus first-run proof.

latest servers

129

Official registry entries marked latest in the fetched window.

github repos

67 (52%)

Listings with a GitHub repository URL attached.

website visible

101 (78%)

Website URL visible in registry metadata, GitHub homepage, or README.

recent movement

62 (48%)

Latest metadata updated within the last 45 days.

package path

20 (16%)

Listings with package install metadata.

remote path

113 (88%)

Listings with hosted remote server metadata.

current risk surface

STDIO is a command boundary, so the public docs need to say what runs.

stdio package path

18 (14%)

Latest listings whose package metadata uses STDIO transport.

stdio docs visible

6 / 12

Readable STDIO package READMEs that mention the command or local-process boundary.

package secrets

11 (9%)

Listings whose package metadata declares secret environment variables.

remote secret headers

24 (19%)

Listings whose remote metadata declares secret request headers.

why now

Local MCP setup is now part of the supply chain.

Recent MCP security research has put STDIO install configs, command review, and package pinning into the trust conversation. A clean launch page should make those boundaries obvious before a user connects a client.

read OX advisory

sales angle

This is a narrow paid review people can understand.

The fix is practical: align registry metadata, README install commands, STDIO warnings, secret handling, smoke tests, and directory proof. That maps directly to Shipcheck and the fixed MCP launch review.

paid MCP review

surface gaps

Common public proof gaps in the current registry sample.

Short registry descriptions 57 / 129 - harder to understand quickly
Missing human-readable title 41 / 129 - namespace carries the listing
Remote-only with no package path 107 / 129 - hosted flow must explain trust and auth clearly
README includes copyable mcpServers config 29 / 47 - among readable GitHub READMEs
README mentions install or client command 24 / 47 - among readable GitHub READMEs
README mentions permissions or safety 37 / 47 - among readable GitHub READMEs
STDIO package READMEs explain command boundary 6 / 12 - among readable STDIO package repos
README mentions server.json 8 / 47 - among readable GitHub READMEs
README includes smoke-test language 12 / 47 - among readable GitHub READMEs

server samples

Searchable public-proof gaps from the fetched registry window.

scope a review

score 42/100 · remote

Agent Skills Search Server

Search and discover Agent Skills from the skills.sh registry. Powered by HAPI MCP server.

  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add install or client command
  • Add smoke-test or tool-list proof

score 48/100 · remote

ac.tandem/docs-mcp

Remote MCP server for Tandem docs, install guides, SDKs, workflows, and agent setup help.

  • Add a human-readable title
  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add install or client command

score 48/100 · package · stdio

ai.aliengiraffe/spotdb

Ephemeral data sandbox for AI workflows with guardrails and security

  • Add a human-readable title
  • Expand the registry description
  • Add smoke-test or tool-list proof
  • Explain STDIO command boundary

score 49/100 · remote

ai.filegraph/document-processing

Extract text from documents, manipulate PDFs, and perform OCR on images.

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Expose a product or docs website

score 54/100 · remote

ai.aient/mcp

MCP-native AI SRE: ask what's broken in production, get a reviewed GitHub fix PR.

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Expose a product or docs website

score 55/100 · remote

ai.explorium/mcp-explorium

Access live company and contact data from Explorium's AgentSource B2B platform.

  • Add a human-readable title
  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add smoke-test or tool-list proof

score 55/100 · package · stdio

SaC — Software as Content

Give your AI agent the ability to respond with live, interactive apps that evolve.

  • Expand the registry description
  • Add install or client command
  • Add smoke-test or tool-list proof
  • Document permissions and safety notes

score 55/100 · remote

ai.adramp/google-ads

Google Ads MCP server — manage campaigns, keywords, and metrics.

  • Attach a public repository
  • Add a human-readable title
  • Expand the registry description
  • Clarify remote auth and trust boundary

score 55/100 · remote

Bowmark

Pre-computed navigation recipes for public websites — skip explore-and-discover.

  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add smoke-test or tool-list proof
  • Document permissions and safety notes

score 55/100 · remote

ai.childanxiety/library

Clinician-reviewed library on anxiety, OCD, and phobias in children ages 5–12.

  • Attach a public repository
  • Add a human-readable title
  • Expand the registry description
  • Clarify remote auth and trust boundary

score 57/100 · remote

Compeller

Create and track AI music videos and audio-reactive visuals from songs.

  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add install or client command
  • Add smoke-test or tool-list proof

score 57/100 · remote

agency.lona/trading

AI-powered trading strategy development: backtesting, market data, and portfolio analysis

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Clarify remote auth and trust boundary

score 57/100 · package · stdio

ai.autoblocks/contextlayer-mcp

Personal context management for AI assistants

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Expose a product or docs website

score 57/100 · package · stdio

ai.autoblocks/ctxl

Personal context management for AI assistants

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Expose a product or docs website

score 57/100 · package · stdio

ai.autoblocks/ctxl-mcp

Personal context management for AI assistants

  • Add a human-readable title
  • Expand the registry description
  • Make README proof available
  • Expose a product or docs website

score 60/100 · package · stdio

ai.adeu/adeu

Automated DOCX Redlining Engine

  • Add a human-readable title
  • Expand the registry description
  • Add smoke-test or tool-list proof
  • Explain STDIO command boundary

score 60/100 · remote

ai.byteask/embedded-docs

Page-cited retrieval for embedded docs, datasheets, MISRA, CMSIS, and RTOS references.

  • Add a human-readable title
  • Expand the registry description
  • Clarify remote auth and trust boundary
  • Add smoke-test or tool-list proof

score 60/100 · remote

Contabo (VPS) MCP Server

Contabo API (v1.0.0) as MCP tools for cloud provisioning, and management. Powered by HAPI MCP server

  • Clarify remote auth and trust boundary
  • Add install or client command
  • Add smoke-test or tool-list proof
  • Add copyable mcpServers config

interpretation

What this means for maintainers.

open checklist

01

Remote servers need more trust copy, not less.

When there is no local package path, the public docs need to make auth, transport, permissions, data access, and support boundaries easy to inspect.

02

Registry metadata should mirror the README's best proof.

A strong README still loses value if the registry title, website, package path, and description do not surface the same clarity.

03

Directory acceptance is becoming a launch system.

The strongest submissions create a proof trail: package, registry metadata, smoke test, safety notes, score page, and curated PR context.

04

Small docs fixes are commercial leverage.

Clear install paths reduce support load, make buyers more comfortable, and give directory maintainers fewer reasons to skip a listing.

sources and method

Public metadata, conservative scoring.

registry

Official MCP Registry API

Fetched 300 registry entries from the public API and kept entries whose official metadata marked them latest.

registry endpoint

docs

MCP Registry docs

The public registry flow uses server metadata for discovery. This pulse treats metadata clarity as part of launch readiness.

registry quickstart

security

MCP STDIO research

Recent public research called out STDIO-driven command execution paths across MCP-adjacent tooling. This pulse treats local command boundaries as public launch proof.

OX advisory

github

Linked GitHub repos

When a listing exposed a GitHub repo, the script read public repo metadata and README text. Private code and private docs were not inspected.

aggregate JSON

shipcheck

Repeatable review path

Shipcheck turns the same launch-readiness ideas into a local scanner, GitHub Action, MCP server, and paid review workflow.

open Shipcheck

use the pulse

Run the free audit before submitting a server.

Paste your metadata and public repo, then use the checklist or a fixed review to close the gaps.

run free audit