@sanctuary-framework/mcp-server 0.7.0 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -156,6 +156,16 @@ Environment variables:
156
156
  | `SANCTUARY_WEBHOOK_SECRET` | HMAC-SHA256 shared secret | — |
157
157
  | `SANCTUARY_WEBHOOK_CALLBACK_PORT` | Callback listener port | `3502` |
158
158
 
159
+ ## Running Alongside Another MCP Server
160
+
161
+ Sanctuary is designed to run as a parallel MCP server — it adds sovereignty infrastructure to your agent without replacing any of its existing tools. Both servers appear in the same session as independent tool providers.
162
+
163
+ For the full setup guide (installation options, passphrase management, systemd units), see [`docs/DEPLOYMENT.md`](docs/DEPLOYMENT.md).
164
+
165
+ For a reference MCP config, see [`docs/examples/parallel-mcp-config.json`](docs/examples/parallel-mcp-config.json).
166
+
167
+ For always-on agents with latency constraints, use the `persistent-agent` Principal Policy template which auto-allows routine operations and only gates destructive actions. See [`src/principal-policy/templates/persistent-agent.yaml`](src/principal-policy/templates/persistent-agent.yaml).
168
+
159
169
  ## Principal Policy (prompt injection defense)
160
170
 
161
171
  The Principal Policy is the human-controlled, agent-immutable configuration that gates operations through a three-tier approval system. It sits between the MCP router and every tool handler — no tool call can bypass it.