Provides identity, governance, and scoped access control for AI agents across various MCP clients.
Also known as proofable/mcp, mcp, @proofable/mcp
- Stars
- 1
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 0.1.10
- Last commit
- Nov 13, 2026
- License
- Apache-2.0
What is Proofable MCP?
Proofable provides identity, governance, and scoped access control for AI agents across various MCP clients.
Summary generated from the project's README
About Proofable MCP server
Proofable MCP provides identity, governance, and scoped access control for AI agents. It allows developers to connect a single profile to various MCP clients including Cursor, Claude, Codex, VS Code, and ChatGPT. By adding the remote URL to a client configuration or using the provided SDK setup command, users can ensure their agents operate within defined permissions and use verified proofs.
This server functions through a set of specialized tools designed to manage agent context and verification workflows. It includes tools like proofable_context for loading profiles, proofable_verify_or_guide for reusing existing proofs, and proofable_agent_link for managing permissions. Users can secure their connection by using a PROOFABLE_ACCESS_KEY as a Bearer token in their server configuration.
Instead of starting every verification from scratch, the system attempts to reuse current proofs and returns statuses such as Passed, Action needed, or Blocked. This creates a structured way to handle identity and tool access across different AI environments without duplicating connections.
What you can do with it
- 1Retrieve user identity and contextAsk the agent to display your current Proofable profile and available proofs.
- 2Verify task requirementsDetermine if a necessary proof is already available for a specific task to avoid redundant steps.
- 3Manage agent permissionsCreate and load specific permissions to control what an agent is allowed to do.
- 4Secure tool executionImplement a permission decision before the host code executes a specific tool.
Written from the project's README. Every use case is checked against a quote from it.
How to install Proofable MCP
Formats verified 10 Nov 2026Install Proofable MCP in Claude Code
$ claude mcp add mcp -- npx -y @proofable/mcpInstall Proofable MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"mcp": {
"command": "npx",
"args": [
"-y",
"@proofable/mcp"
]
}
}
}Install Proofable MCP in VS Code
.vscode/mcp.json
{
"servers": {
"mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@proofable/mcp"
]
}
}
}Install Proofable MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"mcp": {
"command": "npx",
"args": [
"-y",
"@proofable/mcp"
]
}
}
}Install Proofable MCP in Codex
~/.codex/config.toml
[mcp_servers.mcp] command = "npx" args = ["-y", "@proofable/mcp"]
Install Proofable MCP in Gemini CLI
~/.gemini/settings.json
{
"mcpServers": {
"mcp": {
"command": "npx",
"args": [
"-y",
"@proofable/mcp"
]
}
}
}Automated checks
Ran Nov 14, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachableproofable/mcp is public and reachablegithub.com/proofable/mcp ↗Pass
- Package names this repository as its source@proofable/mcp names proofable/mcp as sourcenpmjs.com/package/@proofable/mcp ↗Pass
- Build provenance (Sigstore attestation)Sigstore attestation for @proofable/mcpnpmjs.com/package/@proofable/mcp/v/0.1.10#provenance ↗Pass
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-13github.com/proofable/mcp ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNo popular package one edit away from @proofable/mcpPass
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeHTTPS endpoint answered 401 (sign-in required)Pass
README
View on GitHubProofable MCP
The trust harness for AI. Identity and governance for the agents you run, in every MCP client.
Connect Proofable once, and your agents can use your profile, proofs, permissions, and protected tools in Cursor, Claude, Codex, VS Code, ChatGPT, and any other MCP client.
Start
Open Set up Proofable, choose your client, and finish sign-in. Then ask:
Show my Proofable profile and the proofs I can reuse.
For a specific requirement:
Check whether I already have the proof needed for this task. Reuse it if it qualifies; otherwise show me the next step.
Connect
Add https://mcp.proofable.me/mcp to any MCP client, or use a standard MCP configuration:
{
"mcpServers": {
"proofable": {
"type": "http",
"url": "https://mcp.proofable.me/mcp"
}
}
}
The installer writes the same connection for supported clients:
npx -y @proofable/sdk setup
Claude Code installs the server and the workflow skills together:
/plugin marketplace add proofable/mcp
/plugin install proofable-mcp@proofable
Keep one Proofable connection per client. A marketplace plugin plus a manual entry creates a duplicate.
What agents can do
| Outcome | Tools |
|---|---|
| Load the signed-in profile and current context | proofable_context |
| Find supported checks and their required inputs | proofable_verifiers_catalog |
| Check whether a qualifying proof already exists | proofable_proofs_check |
| Reuse a current proof or get the next verification step | proofable_verify_or_guide |
| Create or refresh a proof | proofable_verify |
| Prepare and submit a check needing another signer | proofable_verify_prepare, proofable_verify_submit |
| Find and read proofs | proofable_proofs_get |
| Update proof metadata or add feedback | proofable_proofs_update |
| Check, create, and load agent permissions | proofable_agent_link, proofable_agent_create, proofable_agent_mount |
Call proofable_context once after connecting. Proofable reuses current proofs before starting another verification and returns Passed, Action needed, or Blocked.
The MCP tool reference lists every tool with its inputs and result shapes.
Servers and automation
Create an access key under Account → Access keys, store it as PROOFABLE_ACCESS_KEY, and send it as a Bearer token. Keep access keys out of source code, browser bundles, and chat.
{
"mcpServers": {
"proofable": {
"type": "http",
"url": "https://mcp.proofable.me/mcp",
"headers": {
"Authorization": "Bearer ${PROOFABLE_ACCESS_KEY}"
}
}
}
}
Examples
- Prompt injection test: hides AgentDojo attacks in a note your agent reads, runs the task under a scoped and a broad permission, and grades every run from its signed receipt.
- Guarded agent action: makes one permission decision before host code runs a tool.
Build with Proofable
Use the Proofable SDK for application code, verification gates, and the CLI. This repository carries the MCP registry manifest, the host plugins, and the workflow skills; the server runs at https://mcp.proofable.me/mcp.
Support
Apache-2.0. Proofable is published by NEUS Network, Inc.
Proofable MCP questions
How do I add Proofable MCP to Claude Code?
Run claude mcp add mcp -- npx -y @proofable/mcp in your terminal. Add --scope user to make it available in every project.
How do I add Proofable MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What happens if I use both a marketplace plugin and a manual entry?
Using both a marketplace plugin and a manual entry will create a duplicate connection.
How can I use Proofable in my own application code?
You can use the Proofable SDK for application code, verification gates, and the CLI.
How do I authenticate my server connection?
Create an access key in your account settings, store it as PROOFABLE_ACCESS_KEY, and send it as a Bearer token in the headers.