fix: transport type (checkbox) & local init not required for commands

Author: divshekhar

.syrin/.dev-history

@@ -2,9 +2,10 @@ Fetch all the network codes
 Give me report for today (25-12-2025) for Rtbdemand
 List all the network codes
 /tools
-what can you do?
 what is today's date?
 inventory?
 give me inventory list
 /clear
 /exit
+what can you do?
+what are the network codes?

README.md

@@ -127,6 +127,8 @@ Requirements: Node.js >= 20.12, npm >= 9
 | `syrin doctor`  | Environment check — validate config and connections      |
 | `syrin list`    | Inspect tools, resources, and prompts from your server   |
 
+**Zero-config:** For `syrin test --connection`, `syrin list`, and `syrin analyse`, no config file is required when you pass `--url <url>` (or `--script <command>` for stdio). `syrin dev` needs local or global config for LLM credentials.
+
 ---
 
 ## Global Configuration

docs/Commands/syrin-config.md

@@ -387,7 +387,7 @@ llm:
 
 **Key Differences from Local Config:**
 
-- No `transport`, `mcp_url`, or `script` fields
+- No `transport`, `url`, or `script` fields
 - `project_name` is always `GlobalSyrin`
 - Focuses only on LLM provider configuration
 

docs/Commands/syrin-doctor.md

@@ -182,7 +182,7 @@ Resolution
 
 - Ensure `transport` matches the configured fields
 - Use `script` for `stdio`
-- Use `mcp_url` for `http`
+- Use `url` for `http`
 
 ## Exit Codes
 

docs/Commands/syrin-init.md

@@ -172,7 +172,7 @@ agent_name: 'My Agent'
 
 transport: 'stdio' # or "http"
 
-mcp_url: 'http://localhost:3000' # required for http
+url: 'http://localhost:3000' # required for http
 script: 'python server.py' # required for stdio
 
 llm:

docs/Configuration/syrin.yaml.md

@@ -35,7 +35,7 @@ agent_name: 'My Agent'
 
 transport: 'stdio' # or "http"
 
-mcp_url: 'http://localhost:3000' # Required for http transport
+url: 'http://localhost:3000' # Required for http transport
 script: 'python server.py' # Required for stdio transport
 
 llm:
@@ -65,7 +65,7 @@ Changes take effect immediately.
 | `project_name` | string                | Yes         | Logical project identifier                |
 | `agent_name`   | string                | Yes         | Logical agent name used in execution      |
 | `transport`    | `"stdio"` or `"http"` | Yes         | MCP transport type                        |
-| `mcp_url`      | string                | Conditional | Required when `transport` is `http`       |
+| `url`          | string                | Conditional | Required when `transport` is `http`       |
 | `script`       | string                | Conditional | Required when `transport` is `stdio`      |
 | `llm`          | object                | Yes         | LLM provider definitions                  |
 
@@ -101,7 +101,7 @@ Use `http` when the MCP server exposes an HTTP endpoint.
 
 ```yaml
 transport: 'http'
-mcp_url: 'http://localhost:3000'
+url: 'http://localhost:3000'
 ```
 
 The server must already be running.\

docs/Configuration/transport-types.md

@@ -110,7 +110,7 @@ In this mode, **Syrin does not own the server lifecycle**.
 
 ```yaml
 transport: 'http'
-mcp_url: 'http://localhost:3000'
+url: 'http://localhost:3000'
 ```
 
 ### How It Works
@@ -139,25 +139,25 @@ This transport is common in **production deployments**.
 Local server:
 
 ```yaml
-mcp_url: 'http://localhost:3000'
+url: 'http://localhost:3000'
 ```
 
 Remote server:
 
 ```yaml
-mcp_url: 'https://api.example.com/mcp'
+url: 'https://api.example.com/mcp'
 ```
 
 With custom path:
 
 ```yaml
-mcp_url: 'http://localhost:3000/mcp/v1'
+url: 'http://localhost:3000/mcp/v1'
 ```
 
 With authentication:
 
 ```yaml
-mcp_url: 'https://user:pass@api.example.com/mcp'
+url: 'https://user:pass@api.example.com/mcp'
 ```
 
 ### Behaviour in Syrin Commands
@@ -207,7 +207,7 @@ The transport choice affects **how much execution control Syrin can exercise**.
 
 ```yaml
 transport: 'http'
-mcp_url: 'http://localhost:3000'
+url: 'http://localhost:3000'
 ```
 
 2. Start the HTTP server

docs/Testing/troubleshooting.md

@@ -366,7 +366,7 @@ Cannot connect to MCP server
 
    ```yaml
    transport: 'http'
-   mcp_url: 'http://localhost:3000' # Correct URL
+   url: 'http://localhost:3000' # Correct URL
    ```
 
 3. **Test Server Manually**

examples/demo-mcp-py/README.md

@@ -75,7 +75,7 @@ This will automatically spawn the server using the script in syrin.yaml.
 
 1. Update `syrin.yaml`:
    - Change `transport: 'http'`
-   - Uncomment and set `mcp_url: 'http://localhost:8000/mcp'`
+   - Uncomment and set `url: 'http://localhost:8000/mcp'`
    - Update `script: '.venv/bin/python server.py --mode http'`
 
 2. Option A: Spawn server internally

examples/demo-mcp-py/syrin.yaml

@@ -25,7 +25,7 @@ transport: "http"
 # MCP Server URL (required for http transport)
 # The HTTP endpoint where your MCP server is running
 # Example: "http://localhost:8000/mcp"
-mcp_url: "http://localhost:8000/mcp"
+url: "http://localhost:8000/mcp"
 
 # Script Configuration
 # --------------------
@@ -38,24 +38,18 @@ script: "python3 server.py"
 # LLM Provider Configuration
 # ---------------------------
 # Configure one or more LLM providers for Syrin to use.
-# API keys and model names can be set as:
-#   - Environment variable names (recommended for security)
-#   - Direct values (not recommended for production)
+# Use environment variable names (keys in .env) for API_KEY and MODEL_NAME.
 # At least one provider must be marked as default: true
 
 llm:
 
   # openai Provider
   openai:
 
-    # API key: Set as environment variable name or direct value
-    # Example: "OPENAI_API_KEY" (reads from process.env.OPENAI_API_KEY)
-    # Or: "sk-..." (direct value, not recommended)
+    # API key: env var name in .env (e.g. OPENAI_API_KEY)
     API_KEY: "OPENAI_API_KEY"
 
-    # Model name: Set as environment variable name or direct value
-    # Example: "OPENAI_MODEL_NAME" (reads from process.env.OPENAI_MODEL_NAME)
-    # Or: "gpt-4" (direct value)
+    # Model name: env var name in .env (e.g. OPENAI_MODEL_NAME)
     MODEL_NAME: "OPENAI_MODEL_NAME"
 
     # Set as default LLM provider
@@ -69,10 +63,10 @@ llm:
 
   # Ollama Provider - uncomment to enable
   # ollama:
-  #   MODEL_NAME: "OLLAMA_MODEL_NAME"  # or "llama2", "mistral", etc.
+  #   MODEL_NAME: "OLLAMA_MODEL_NAME"
   #   default: false
 
-# Tool Validation Configuration (v1.3.0)
+# Tool Validation Configuration
 # ---------------------------
 # Configure tool-level structural safety validation
 # This section is optional but recommended for production use