mcpField
F
firmware-attestation-mcp

firmware-attestation-mcp

Local
github.com/CSOAI-ORGSecurity0Updated 19w ago

About firmware-attestation-mcp

firmware-attestation-mcp checks a host’s firmware trust state and can gate AI inference on the result. It reports risk indicators rather than declaring a system clean.

Key features

  1. 01Read-only firmware evidence scan
  2. 02Check persistence-implant indicators
  3. 03Create HMAC-signed attestations
  4. 04Allow or block inference
  5. 05List defended attack surfaces

Use cases

  1. 01Review host firmware trust evidence
  2. 02Check defenses against persistence threats
  3. 03Require verified trust before inference

MCP support

It exposes firmware scanning, signature checks, attestation, inference gating, and threat-model tools to MCP clients. Attestations can be verified at `proofof.ai/api/verify`.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"csoai-org-firmware-attestation": {
"command": "uvx",
"args": [
"firmware-attestation-mcp"
]
}
}
}
Run commanduvx firmware-attestation-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypifirmware-attestation-mcp1.0.1

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

MCP Scorecard: 88/100

Firmware Attestation MCP

Hardware trust layer for sovereign AI. Persistence implants live below the OS (BIOS/UEFI, SMM, network boot ROMs, HDD HPA) and survive OS reinstalls and disk wipes. This MCP attests a host's firmware trust state and gates inference on a verified result.

Tools

ToolWhat
scan_firmwareread-only host evidence (Secure Boot, TPM, SIP, BIOS, HPA)
check_ant_signaturesmatch to NSA-ANT-class persistence preconditions + defenses
attest_firmwareHMAC-signed attestation, verifiable at proofof.ai/api/verify
gate_inferenceALLOW/BLOCK AI on this host (strict by default)
list_threat_modelthe attack surface this defends against

Honest by design

Reports indicators (preconditions implants rely on), never "clean". A BLOCK means "lacks confirmed trust anchors," not "hacked." Harden per the listed defenses, then re-gate.

pip install firmware-attestation-mcp

© CSOAI LTD (trading as MEOK AI Labs) · MIT

Configuration

Add to your claude_desktop_config.json (Claude Desktop) or your MCP client config:

{
  "mcpServers": {
    "firmware-attestation-mcp": {
      "command": "uvx",
      "args": ["firmware-attestation-mcp"]
    }
  }
}

Or: pip install firmware-attestation-mcp then run the firmware-attestation-mcp command (stdio transport).

Examples

Once configured, ask your assistant, for example:

  • "Use scan_firmware to …"
  • "Use check_ant_signatures to …"
  • "Use attest_firmware to …"

Part of the MEOK constellation

This MCP is one node in a connected ecosystem built by MEOK AI LABS around a single sovereign AI core — governed agents with a hash-chained audit trail, mapped to the CSOAI compliance charter.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.