SkillAgentSearch skills...

lodestone

Use Lodestone (lodestone.studiocorsair.com) as one MCP connector for Meta Ads, Instagram insights, Google Analytics 4, email, Telegram, Typefully, and image/video generation.

Install / Use

npx skills add Anil-matcha/awesome-muse-connectors --skill lodestone

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

85/100

Supported Platforms

Universal

Our assessment of lodestone

lodestone scores 85/100 on our quality scale, 325th of 420 Communication skills we index.

Its SKILL.md is 4.0 KB long, well organised into 11 sections with 1 code example: a solid amount of guidance for an agent.

With 1,321 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
26/30
Structure
17/20
Description
15/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated 20 days ago, so lodestone is actively maintained.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

lodestone compared with similar skills

All 4 of these similar skills score higher than lodestone; compare them before choosing.

SkillScoreStarsUpdatedFormat
lodestone (this skill)by Anil-matcha851.3k20d agoSKILL.md
taste-skillby Leonxlnx100102.3ktodayCLAUDE.md
Agent-Reachby Panniantong100102.2ktodayCLAUDE.md
headroomby headroomlabs-ai10077.8k1d agoCLAUDE.md
CowAgentby zhayujie10047.6ktodayCLAUDE.md

Frequently asked questions

How do I install lodestone?
Run npx skills add Anil-matcha/awesome-muse-connectors --skill lodestone. The install tabs above show the steps for each supported agent.
Which AI agents does lodestone work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is lodestone safe to use?
It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is lodestone still maintained?
The repository was last updated 20 days ago, so lodestone is actively maintained.

name: "lodestone" description: "Use Lodestone (lodestone.studiocorsair.com) as one MCP connector for Meta Ads, Instagram insights, Google Analytics 4, email, Telegram, Typefully, and image/video generation. Trigger phrases: lodestone, meta ads, ad campaign insights, instagram insights, ga4 report, google analytics report, send email, telegram message, typefully post, generate image, generate video." metadata: { "includeInPrompt": true } tagline: "One key for Meta Ads, Instagram insights, GA4, email, Telegram, Typefully, and image/video generation." catalog_auth: "API key (per-user Lodestone connector key)" catalog_hosts: ["lodestone.studiocorsair.com"]

Lodestone

Purpose

Lodestone is a hosted MCP server that puts several services behind one per-user key. Muse should reach for it when a task needs the user's own Meta Ads account, Instagram account insights, Google Analytics 4 reports, email, a Telegram bot, a Typefully queue, or AI image and video generation.

The connector uses the muse-directory connection origin, which exposes 39 tools: 24 read-only and 15 that create, change, spend, or send. Each service beyond the key (Google Analytics, Meta Ads, Typefully, and so on) is connected by the user in their Lodestone dashboard; a tool for an unconnected service replies with the dashboard link instead of failing silently.

Source: the server's own tools/list response (MCP Streamable HTTP, JSON-RPC), checked 2026-10-09. Lodestone's MCP OAuth flow did not complete from Muse in testing, so this connector uses the API key path (Authorization: Bearer).

Disclosure: the contributor built and operates Lodestone.

Tooling

All commands go through bin/lodestone.py (Python standard library only).

# Verify the key and reach the server (prints the tool count)
bin/lodestone.py auth

# List every tool; WRITE marks anything that is not read-only
bin/lodestone.py tools

# Call a read-only tool
bin/lodestone.py call typefully_whoami
bin/lodestone.py call meta_ads_list_accounts
bin/lodestone.py call ga4_run_report --args '{"property_id":"123456789"}'

# Call a tool that creates, changes, spends, or sends (needs explicit user approval first)
bin/lodestone.py call email_send --args '{"to":"a@example.com","subject":"Hi","body":"Hello"}' --confirm

Run bin/lodestone.py tools first. Tool argument names come from the server; call forwards --args as the JSON arguments object unchanged.

Auth

  • Provider id: lodestone (credential is collected as custom.lodestone)
  • Collection: API key via the secure credential flow (credentials.request_api_access). The user copies their connector key from their Lodestone dashboard. For local use, export it as LODESTONE_KEY. The key is never accepted as a command-line argument.
  • Required scopes: none; the key identifies the user's Lodestone account
  • Allowed hosts: lodestone.studiocorsair.com
  • Status check: bin/lodestone.py auth

Operating Rules

  1. call refuses any tool whose server annotations are not readOnlyHint: true unless --confirm is passed. Tools that need it include email_send, telegram_send_message, typefully_post, every meta_ads_create_* / meta_ads_update_* / meta_ads_set_status / meta_ads_boost_post, and the five generate_* media tools.
  2. Before using --confirm, show the user the exact tool and arguments and wait for approval. Never pass --confirm on your own initiative.
  3. Meta Ads writes can spend real money and email, Telegram, and Typefully writes reach real people. Prefer drafts and read-only calls.
  4. The key is a secret. Never print, log, or commit it.
  5. Only lodestone.studiocorsair.com is contacted.

Files

  • SKILL.md
  • bin/lodestone.py

Maturity

✅ Live-tested against the real server on 2026-10-09 with a real connector key: auth (39 tools), tools (24 read, 15 write), call on read-only tools, call refusing email_send without --confirm, unknown tool, bad key (HTTP 401), and missing key. The --confirm write path was deliberately not run live because every write tool sends, spends, or publishes.

Related Skills

View on GitHub
GitHub Stars1.3k
CategoryCommunication
Updated20d ago
Forks305

Languages

Python

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions