Access levels of skills and MCP servers

Broad access

• AI Builder Club SkillsPlugin
  • The skills run, test and ship code autonomously
  • They spin up cloud boxes and delegate work to other CLI agents, running shell commands and scripts
• AI Code InterfaceCLI
  • Starts agent CLIs that edit files and run commands, and the package itself runs generated shell commands after your actions
  • Passes code, diffs, git history and screenshots from the editor to agents
  • Automatic test runs after edits are switched on by the ai-code-auto-test-type setting
• AI DevKitCLI
  • Sends instructions to running agents, including remotely from Telegram and Slack
  • Changes agent and MCP settings in the project
• AI EmployeesPlugin
  • Employees drive your browser and PC inside your signed-in accounts
  • Scheduled routines read your inbox, ad accounts and customer data and prepare actions in your name
  • The last click stays yours only until you release a channel for autonomous work
• AI Finansist: MCP server for 1CMCP server
  • The OData client can create, update and delete 1C database records, not just read them
  • The 1C login and password are stored in .env in plain text
• AI Game Developer (Unity MCP)MCP server
  • Executes arbitrary C# code in the editor
  • Can delete project assets, scripts and packages
• AI Game Developer: Unreal MCPMCP server
  • The agent changes actors, assets, Blueprints and C++ code of the project, and rql-mutate mutations touch editor data
  • Cloud mode sends editor commands and data to the ai-game.dev service
• AI Gateway LabsWorkflow
  • Create paid Azure resources
  • Contributor and RBAC Administrator rights are needed
• AI MaestroCLI
  • Runs agents with terminal and repository access
  • Opens network connections between machines
  • Installs via a script from the internet
• AI Search HubSkill
  • Starts a browser with remote debugging and drives your logged-in sessions on third-party platforms
  • Seeds a separate debug profile from your browser data
  • Automating these platforms may violate their terms of use
• AI-EDTPlugin
  • Changes configuration metadata and code and updates the connected infobase
  • Controls the debugger and reads infobase data
• AIO SandboxMCP server
  • Executes arbitrary shell commands and code inside the container on the agent's request
  • Without a configured access key the container's services are unauthenticated
  • Gives the agent control over the sandbox's browser and filesystem
• AIPexMCP server
  • It controls your real browser with open sessions
  • It can perform on-page actions and read page content
• ANUSCLI
  • The agent executes shell commands and edits project files, and a line starting with `!` runs a command directly
  • File contents and prompts go to free providers, some of which may use them for training
  • Provider keys sit in the `auth.json` file under `~/.anus/agent`
• AOCI-CODEMCP server
  • Writes index files, agent rules and an `.aoci/` directory into the repository, and with `--hooks` edits the agent settings
  • Connects to the database you declare through a connection string from an environment variable and reads table metadata
  • Starts background processes: the MCP server and a local panel on a loopback address
• ARISPlugin
  • Autonomously runs experiments and executes code with minimal supervision
  • Calls external models to review steps
• ARL-NextMCP server
  • The ARL token opens the company's whole database of assets, vulnerabilities and found leaks, and replies land in the model's context
  • The platform itself scans external targets with Nmap and Nuclei, and unauthorized scanning of other people's systems is illegal
  • The ARL_TLS_VERIFY=false variable disables certificate checks and allows connection interception
• ATLAS MCP ServerMCP server
  • Creates, changes and deletes projects, tasks and knowledge in your Neo4j database
  • The atlas_database_clean tool wipes the whole database in one call
  • Keeps the Neo4j password in the client config or in .env
• AWS Skills for Claude CodePlugin
  • Uses AWS credentials and can run aws commands
  • Cloud actions can change infrastructure and costs
• ActionbookMCP server
  • The agent clicks, types and runs JavaScript in your signed-in sessions, and actions taken on sites in your name cannot be undone
  • The extension reads and changes cookies and asks for access to all addresses, so data from gated pages reaches the agent
  • Page content and commands pass through the hosted server edge.actionbook.app, and an outside service, Clerk, handles sign-in
• AdLoopMCP server
  • Writes to live Google Ads and Reddit Ads accounts and manages campaign budgets
  • OAuth grants access to conversion data, search terms and Tag Manager and Search Console settings
  • In the hosted edition, tokens and history are stored on AdLoop Cloud servers in Germany
• Ads MCP on Cloudflare Workers: Direct and VK AdsMCP server
  • The generic direct_request and vk_ads_request tools can call any API method, including changing campaigns
  • Ad tokens are stored as Cloudflare Workers secrets
  • Server access is protected only by a secret in the URL path
• Agenova integration catalogMCP server
  • May send messages and actions in chat on request without a separate confirmation screen
  • Works with seller tokens and CRM via a third-party hosted dashboard
• Agent ApprenticeshipCLI
  • Launches local agent CLIs that run commands and change files in the working folder
  • Stores model provider keys in ~/.agent-apprenticeship/.env.local and sends task context to those providers
  • In the public mode shares learning signals with the shared ecosystem
• Agent Client for ObsidianPlugin
  • Launches agents as child processes with full terminal and system access, same as a plain CLI
• Agent DeckCLI
  • It starts, restarts and closes agent sessions and sends them messages, and worktree scripts from the repository run as you after approval
  • The conductor mode accepts commands from Telegram or Slack through a background bridge, so sessions can be controlled from a phone
  • Web mode puts session terminals in a browser on port 8420, with access protected by a token
• Agent KanbanSkill
  • The skills let an agent create, assign and cancel board tasks and submit work using Realmroot tokens
  • The worker agent edits the repository and runs git push and gh commands through Realmroot
  • The maintainer agent acts on a schedule with no human present at launch
• Agent ScriptsSkill
  • `sync-skills` changes links in `~/.codex` and `~/.claude`, and the shared `AGENTS.MD` sets rules for all your agent tasks
  • The set includes skills that send messages, buy numbers and text through Twilio, change DNS and deploy code with Wrangler
  • `browser-use` works in your Chrome with its open sessions, and several skills read messenger conversations
• Agent Skills for Google productsSkill
  • The skills run gcloud commands that change cloud infrastructure and IAM permissions
  • Actions in Google Cloud may incur charges and affect production projects
• Agent StuffSkill
  • The skills read Apple Mail, call the Google Workspace and Sentry APIs and work with local keys
  • The extensions replace the bash tool, run child processes in tmux and can trust repositories automatically
  • The web-browser skill controls Chrome, while ghidra and openscad launch external programs
• Agent Toolkit for AWSPlugin
  • The MCP Server calls your AWS account API and runs scripts
  • It can create and change cloud resources
• AgentBridge: a bridge between Claude Code and CodexPlugin
  • The launch commands switch off permission prompts for both agents by default, so they run commands and edit files without your confirmation
  • A background process listens on local ports from 4500 and relays messages between the agents
  • One agent's message content flows into the other agent's thread, so instructions inside it may be acted on
• AgentCall join-meetingSkill
  • It connects to an external service and joins live meetings on your behalf
  • It can create an AgentCall account and spend credits, sending voice and screenshots to the cloud
• AgentDockMCP server
  • Runs commands and edits files on real machines and servers
  • Exposes network access to the runtime with a token
• AgentSysPlugin
  • Creates branches and PRs and merges in the remote repository
  • Runs shell commands and multiple agents
• AgenterCLI
  • Skills directly modify configuration XML metadata and write to the 1C infobase via 1cv8.exe
  • Needs an Anthropic API key and 1C admin database credentials in local files
• AgentgatewayMCP server
  • Sits in the path of all agent traffic to MCP servers, LLMs and other agents
  • Configures authentication and access policies; a misconfiguration can open up unintended access
• Agentic QE FleetPlugin
  • Agents run tests and arbitrary commands in the project
  • Chaos and pentest agents can inject faults and probe test environments
  • Creates a local database and agent configuration in the project
• AgenticaCLI
  • The agent writes and edits files and runs shell commands, with the sandbox enabled by the --sandbox flag
  • The gateway opens a web interface and takes instructions from messengers, so protect it with a password and do not expose the port
  • Provider keys, sessions and memory sit in ~/.agentica and part of the context goes to the chosen model provider
• AgenvoyMCP server
  • Runs as a persistent local daemon with access to the user's files, schedules and memory, plus a web dashboard on a local port
  • Stores external MCP servers' OAuth tokens in the OS keychain and builds new tools itself without approving each step
• Aident LoadoutMCP server
  • Through connected accounts the agent reads and sends email, posts to Slack and Notion and deletes records: Gmail's 61 actions include permanent deletion that bypasses the trash
  • Aident Vault keeps the tokens of connected services on Aident AI, LLC's side, and task text and service responses pass through the company's servers
  • Paid actions deduct credits from the Aident balance, so set a spending approval threshold before the first task
• AiderCLI
  • Edits files and commits automatically
  • Can run linters, tests and shell commands
  • Code is sent to the model provider
• AionUiCLI
  • The agent reads and writes files on your computer
  • Stores model provider API keys
• Aiven MCP ServerMCP server
  • Can create, change and delete cloud services and data
  • Runs write SQL and produces Kafka messages
• Alfa API MCP: Alfa-Bank accounts and paymentsMCP server
  • create_payment_order can move real money once a signing module is connected
  • Requires mTLS certificates and a bank agreement; a misconfiguration can expose banking data
• Alice Voice for HermesPlugin
  • The secret URL plays the role of a password and the Yandex signature is not verified
  • Text becomes an agent command with the rights of the Telegram session
  • The integration is experimental, for one operator
• Aliyun Model Studio CLICLI
  • Stores the Model Studio API key and, when needed, an Alibaba Cloud AccessKey that can manage subscription seats
  • Fine-tuning, model deployment and generation calls spend money on the platform account
  • Installing through a curl or PowerShell script and updating with bl update run code fetched from the network
• AlookCLI
  • People in a room message an agent that runs commands on your machine with your permissions
  • A daemon keeps a persistent WebSocket link to the hosted part at alook.ai
  • The agent keeps memory and an inbox, so conversation content accumulates
• Alpaca MCP ServerMCP server
  • In live mode it places real orders, with a risk of money loss
  • You must check paper versus live mode before use
  • It requires secret API keys and access to the portfolio
• Altinity MCP ServerMCP server
  • Runs arbitrary SQL against the database, including writes unless read_only is set
  • Holds the ClickHouse password and OAuth or JWE secrets