Minor extension of print_system_prompt command, added explanation on using it to readme [ci skip]

README.md

@@ -354,8 +354,12 @@ For example, when using `uvx`, you would run
 claude mcp add serena -- uvx --from git+https://github.com/oraios/serena serena-mcp-server --context ide-assistant --project $(pwd)
 ```
 
-ℹ️ Once in Claude Code, run `/mcp__serena__initial_instructions` to load instructions for using Serena's tools. Run this command 
-whenever you start a new conversation and after any compacting operation to ensure Claude remains properly configured to use Serena's tools.
+ℹ️ Serena comes with instructions on how to use tools, and Claude needs to read them to properly use serena's tools. 
+  Once in Claude Code, you can ask to "read Serena's initial instructions" or to run `/mcp__serena__initial_instructions` to load instructions for using Serena's tools. 
+  Do this whenever you start a new conversation and after any compacting operation to ensure Claude remains properly configured to use Serena's tools.
+
+ℹ️ **NEW**: an alternative to the above is adding the instructions as part of the system prompt, then you will not need to run the command above or to remember re-running it after compacting.
+  This can be achieved throug starting claude code with `cclaude --append-system-prompt $(uvx --from git+https://github.com/oraios/serena print-system-prompt)`. This is **experimental**, Claude may not understand the instructions correctly in this way. Please report any issues you encounter.
 
 
 ### Claude Desktop

src/serena/agent.py

@@ -127,28 +127,34 @@ def index_project(project: str, log_level: str = "INFO") -> None:
 @click.command()
 @click.argument("project", type=click.Path(exists=True), required=False, default=os.getcwd())
 @click.option("--log-level", type=click.Choice(["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"]), default="WARNING")
-def print_system_prompt(project: str, log_level: str = "WARNING") -> None:
+@click.option(
+    "--only-instructions",
+    is_flag=True,
+    help="If set, only print the initial instructions prompt (without prefix or postfix). The prefix and postfix are small additions that are used to"
+    "make the initial instructions more digestible for an LLM in the context of a system prompt, as opposed to their normal usage within a conversation.",
+)
+def print_system_prompt(project: str, log_level: str = "WARNING", only_instructions: bool = False) -> None:
     """
-    Print the system prompt (initial instructions) for a project.
+    Print the system prompt (initial instructions with a potential pre/post-fix) for a project.
+    """
+    prefix_to_instructions_prompt = (
+        "You will receive access to Serena's symbolic tools. Below are instructions for using them, take them into account."
+    )
+    postfix_to_instructions_prompt = "You begin the conversation by acknowledging that you understood the above instructions on the symbolic tools and are ready to receive a task. You will not need to call the tool on getting initial instructions."
+
+    from serena.tools.workflow_tools import InitialInstructionsTool
 
-    :param project: the project to get the system prompt for. By default, the current working directory is used.
-    """
     log_level_int = logging.getLevelNamesMapping()[log_level.upper()]
     logging.configure(level=log_level_int)
 
-    # Create a SerenaAgent instance
-    agent = SerenaAgent(project=os.path.abspath(project))
-
-    # Get the InitialInstructionsTool instance
-    from serena.tools.workflow_tools import InitialInstructionsTool
-
+    agent = SerenaAgent(project=os.path.abspath(project), serena_config=SerenaConfig(web_dashboard=False, log_level=log_level_int))
     initial_instructions_tool = agent.get_tool(InitialInstructionsTool)
-
-    # Apply the tool to get the system prompt
-    system_prompt = initial_instructions_tool.apply()
-
-    # Print the system prompt
-    print(system_prompt)
+    initial_instructions = initial_instructions_tool.apply()
+    if only_instructions:
+        print(initial_instructions)
+    else:
+        system_prompt = f"{prefix_to_instructions_prompt}\n{initial_instructions}\n{postfix_to_instructions_prompt}"
+        print(system_prompt)
 
 
 class SerenaAgent:

src/serena/tools/workflow_tools.py

@@ -132,8 +132,8 @@ class InitialInstructionsTool(Tool, ToolMarkerDoesNotRequireActiveProject):
     def apply(self) -> str:
         """
         Get the initial instructions for the current coding project.
-        You should always call this tool before starting to work (including using any other tool) on any programming task!
-        The only exception is when a user asks you to activate a project, in which case you should call the `activate_project` first
-        instead and then call this tool.
+        If you haven't received instructions on how to use Serena's tools in the system prompt,
+        you should always call this tool before starting to work (including using any other tool) on any programming task,
+        the only exception being when you are asked to call `activate_project`, which you should then call before.
         """
         return self.agent.create_system_prompt()