- Home
- Skills
- Workflow Automation
- Hermes Bridge
Works with the AI Agents you already use
Hermes Bridge
Let cloud assistants use your self-hosted Hermes Agent over Tailscale, with risky writes gated by a yes/no reply on Telegram.
$2
Hermes Bridge
Example session with this skill installed
Set up hermes-bridge so Claude Code on my laptop can search my Hermes memory.
- Read your context and instructions
- Compiled the hermes bridge
Installs the bridge on your Hermes host bound to its tailnet IP, registers the MCP server, schedules the approval resolver, runs the smoke test, and prints the .mcp.json snippet for Claude Code.
Connects securely to your tools. The creator never sees your data.
About this skill
A thin MCP server that lets a cloud assistant (a tailnet node, Claude Code on another machine, a custom Python client) talk to the user's Hermes Agent over Tailscale. It runs on the Hermes box, binds to a tailnet IP only, and gates risky actions (writes, sends) behind a Telegram approval queue so the user keeps control of what the cloud side can do.
Read-only tools (memory search, vault read, session search, inbox) work
automatically. Write/send tools require the user to reply yes <uuid> or
no <uuid> in their Telegram chat. Tailscale membership is the auth — no
tokens, no API keys.
On the Hermes host, scripts/setup.sh will:
- Detect the user's Tailscale IP and write
~/.hermes/bridge/peer_config.toml(preserved on re-runs) - Copy the bridge files to
~/.hermes/bridge/ - Generate
allowlist.yamlfrom the peer config - Render a systemd user unit at
~/.hermes/bridge/systemd/hermes-bridge.service - Register the bridge as a local MCP server in the user's own Hermes config
- Install helper scripts and create a once-per-minute
hermes-bridge-approvalscron that posts approval requests to Telegram and resolvesyes/noreplies
Requires Hermes Agent and Tailscale; not a generic MCP server. memory_search/session_search take ~10-15 s per call. Approvals expire after 30 min. Restarting the bridge drops calls that are waiting for approval.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago
Requires
Permissions
Allowed Hosts
File Scopes
Binds only to your Tailscale IP; loopback and LAN are refused and non-tailnet clients get 403. No public exposure. Writes and outbound messages need a Telegram 'yes <id>' reply; vault writes are snapshotted; nothing is deleted.