zrb 1.10.2__tar.gz → 1.12.0__tar.gz

{zrb-1.10.2 → zrb-1.12.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: zrb
-Version: 1.10.2
+Version: 1.12.0
 Summary: Your Automation Powerhouse
 Home-page: https://github.com/state-alchemists/zrb
 License: AGPL-3.0-or-later
@@ -27,7 +27,7 @@ Requires-Dist: pdfplumber (>=0.11.6,<0.12.0) ; extra == "rag" or extra == "all"
 Requires-Dist: playwright (>=1.53.0,<2.0.0) ; extra == "playwright" or extra == "all"
 Requires-Dist: prompt-toolkit (>=3.0.51,<4.0.0)
 Requires-Dist: psutil (>=7.0.0,<8.0.0)
-Requires-Dist: pydantic-ai (>=0.3.4,<0.4.0)
+Requires-Dist: pydantic-ai (>=0.4.4,<0.5.0)
 Requires-Dist: pyjwt (>=2.10.1,<3.0.0)
 Requires-Dist: python-dotenv (>=1.1.1,<2.0.0)
 Requires-Dist: python-jose[cryptography] (>=3.4.0,<4.0.0)

{zrb-1.10.2 → zrb-1.12.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "zrb"
-version = "1.10.2"
+version = "1.12.0"
 description = "Your Automation Powerhouse"
 authors = ["Go Frendi Gunawan <gofrendiasgard@gmail.com>"]
 license = "AGPL-3.0-or-later"
@@ -49,7 +49,7 @@ beautifulsoup4 = "^4.13.3"
 fastapi = {extras = ["standard"], version = "^0.115.14"}
 python-jose = {extras = ["cryptography"], version = "^3.4.0"}
 ulid-py = "^1.1.0"
-pydantic-ai = "^0.3.4"
+pydantic-ai = "^0.4.4"
 tiktoken = "^0.8.0"
 openai = {version = "^1.86.0", optional = true}
 playwright = {version = "^1.53.0", optional = true}

{zrb-1.10.2 → zrb-1.12.0}/src/zrb/builtin/llm/chat_session.py

@@ -19,27 +19,29 @@ async def read_user_prompt(ctx: AnyContext) -> str:
     Orchestrates the session by calling helper functions.
     """
     _show_info(ctx)
-    is_web = ctx.env.get("_ZRB_WEB_ENV", "0") == "1"
     final_result = await _handle_initial_message(ctx)
-    if is_web:
+    if ctx.is_web_mode:
         return final_result
-    is_interactive = sys.stdin.isatty()
-    reader = await _setup_input_reader(is_interactive)
+    is_tty = ctx.is_tty
+    reader = await _setup_input_reader(is_tty)
     multiline_mode = False
+    current_modes = ctx.input.modes
     user_inputs = []
     while True:
         await asyncio.sleep(0.01)
         # Get user input based on mode
         if not multiline_mode:
             ctx.print("💬 >>", plain=True)
-        user_input = await _read_next_line(is_interactive, reader, ctx)
+        user_input = await _read_next_line(is_tty, reader, ctx)
         if not multiline_mode:
             ctx.print("", plain=True)
         # Handle user input
         if user_input.strip().lower() in ("/bye", "/quit", "/q", "/exit"):
             user_prompt = "\n".join(user_inputs)
             user_inputs = []
-            result = await _trigger_ask_and_wait_for_result(ctx, user_prompt)
+            result = await _trigger_ask_and_wait_for_result(
+                ctx, user_prompt, current_modes
+            )
             if result is not None:
                 final_result = result
             break
@@ -50,9 +52,18 @@ async def read_user_prompt(ctx: AnyContext) -> str:
             multiline_mode = False
             user_prompt = "\n".join(user_inputs)
             user_inputs = []
-            result = await _trigger_ask_and_wait_for_result(ctx, user_prompt)
+            result = await _trigger_ask_and_wait_for_result(
+                ctx, user_prompt, current_modes
+            )
             if result is not None:
                 final_result = result
+        elif user_input.strip().lower().startswith("/mode"):
+            mode_parts = user_input.split(" ", maxsplit=2)
+            if len(mode_parts) > 1:
+                current_modes = mode_parts[1]
+            ctx.print(f"Current mode: {current_modes}", plain=True)
+            ctx.print("", plain=True)
+            continue
         elif user_input.strip().lower() in ("/help", "/info"):
             _show_info(ctx)
             continue
@@ -62,7 +73,9 @@ async def read_user_prompt(ctx: AnyContext) -> str:
                 continue
             user_prompt = "\n".join(user_inputs)
             user_inputs = []
-            result = await _trigger_ask_and_wait_for_result(ctx, user_prompt)
+            result = await _trigger_ask_and_wait_for_result(
+                ctx, user_prompt, current_modes
+            )
             if result is not None:
                 final_result = result
     return final_result
@@ -75,16 +88,26 @@ def _show_info(ctx: AnyContext):
         ctx: The context object for the task.
     """
     ctx.print(
-        (
-            f"  {stylize_bold_yellow('/bye')}   {stylize_faint('Quit from chat session')}\n"
-            f"  {stylize_bold_yellow('/multi')} {stylize_faint('Start multiline input')}\n"
-            f"  {stylize_bold_yellow('/end')}   {stylize_faint('End multiline input')}\n"
-            f"  {stylize_bold_yellow('/help')}  {stylize_faint('Show this message')}\n"
+        "\n".join(
+            [
+                _format_info_line("/bye", "Quit from chat session"),
+                _format_info_line("/multi", "Start multiline input"),
+                _format_info_line("/end", "End multiline input"),
+                _format_info_line("/modes", "Show current modes"),
+                _format_info_line("/modes <mode1,mode2,..>", "Set current modes"),
+                _format_info_line("/help", "Show this message"),
+            ]
         ),
         plain=True,
     )
 
 
+def _format_info_line(command: str, description: str) -> str:
+    styled_command = stylize_bold_yellow(command.ljust(25))
+    styled_description = stylize_faint(description)
+    return f"  {styled_command} {styled_description}"
+
+
 async def _handle_initial_message(ctx: AnyContext) -> str:
     """Processes the initial message from the command line."""
     if not ctx.input.message or ctx.input.message.strip() == "":
@@ -95,6 +118,7 @@ async def _handle_initial_message(ctx: AnyContext) -> str:
     result = await _trigger_ask_and_wait_for_result(
         ctx,
         user_prompt=ctx.input.message,
+        modes=ctx.input.modes,
         previous_session_name=ctx.input.previous_session,
         start_new=ctx.input.start_new,
     )
@@ -132,6 +156,7 @@ async def _read_next_line(is_interactive: bool, reader, ctx: AnyContext) -> str:
 async def _trigger_ask_and_wait_for_result(
     ctx: AnyContext,
     user_prompt: str,
+    modes: str,
     previous_session_name: str | None = None,
     start_new: bool = False,
 ) -> str | None:
@@ -149,7 +174,7 @@ async def _trigger_ask_and_wait_for_result(
     """
     if user_prompt.strip() == "":
         return None
-    await _trigger_ask(ctx, user_prompt, previous_session_name, start_new)
+    await _trigger_ask(ctx, user_prompt, modes, previous_session_name, start_new)
     result = await _wait_ask_result(ctx)
     md_result = _render_markdown(result) if result is not None else ""
     ctx.print("\n🤖 >>", plain=True)
@@ -194,12 +219,14 @@ def get_llm_ask_input_mapping(callback_ctx: AnyContext):
         "start-new": data.get("start_new"),
         "previous-session": data.get("previous_session_name"),
         "message": data.get("message"),
+        "modes": data.get("modes"),
     }
 
 
 async def _trigger_ask(
     ctx: AnyContext,
     user_prompt: str,
+    modes: str,
     previous_session_name: str | None = None,
     start_new: bool = False,
 ):
@@ -219,6 +246,7 @@ async def _trigger_ask(
             "previous_session_name": previous_session_name,
             "start_new": start_new,
             "message": user_prompt,
+            "modes": modes,
         }
     )
 

{zrb-1.10.2 → zrb-1.12.0}/src/zrb/builtin/llm/llm_ask.py

@@ -65,6 +65,14 @@ _llm_ask_inputs = [
         allow_positional_parsing=False,
         always_prompt=False,
     ),
+    TextInput(
+        "modes",
+        description="Modes",
+        prompt="Modes",
+        default="coding",
+        allow_positional_parsing=False,
+        always_prompt=False,
+    ),
     BoolInput(
         "start-new",
         description="Start new conversation (LLM will forget everything)",
@@ -101,6 +109,9 @@ llm_ask: LLMTask = llm_group.add_task(
         system_prompt=lambda ctx: (
             None if ctx.input.system_prompt.strip() == "" else ctx.input.system_prompt
         ),
+        modes=lambda ctx: (
+            None if ctx.input.modes.strip() == "" else ctx.input.modes.split(",")
+        ),
         message="{ctx.input.message}",
         retries=0,
     ),

{zrb-1.10.2 → zrb-1.12.0}/src/zrb/builtin/llm/tool/file.py

@@ -10,7 +10,7 @@ from zrb.config.llm_rate_limitter import llm_rate_limitter
 from zrb.context.any_context import AnyContext
 from zrb.util.file import read_file, read_file_with_line_numbers, write_file
 
-_EXTRACT_INFO_FROM_FILE_SYSTEM_PROMPT = CFG.LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT
+_EXTRACT_INFO_FROM_FILE_SYSTEM_PROMPT = CFG.LLM_FILE_EXTRACTOR_SYSTEM_PROMPT
 
 
 DEFAULT_EXCLUDED_PATTERNS = [
@@ -471,7 +471,7 @@ async def analyze_file(
     _analyze_file = create_sub_agent_tool(
         tool_name="analyze_file",
         tool_description="analyze file with LLM capability",
-        system_prompt=CFG.LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT,
+        system_prompt=CFG.LLM_FILE_EXTRACTOR_SYSTEM_PROMPT,
         tools=[read_from_file, search_files],
     )
     payload = json.dumps(

{zrb-1.10.2 → zrb-1.12.0}/src/zrb/config/config.py

@@ -19,77 +19,20 @@ Your Automation Powerhouse
 🐤 Follow us at: https://twitter.com/zarubastalchmst
 """
 
-_DEFAULT_LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT = (
-    "You are an intelligent code and configuration analysis agent.\n"
-    "Your primary goal is to extract key information from the provided file(s) "
-    "that is directly relevant to the main assistant's objective.\n"
-    "\n"
-    "Analyze the file content and determine its type (e.g., Python script, "
-    "YAML configuration, Dockerfile, Markdown documentation).\n"
-    "Based on the file type, extract the most important information in a "
-    "structured markdown format.\n"
-    "\n"
-    "- For source code (e.g., .py, .js, .go): Extract key components like "
-    "classes, functions, important variables, and their purposes.\n"
-    "- For configuration files (e.g., .yaml, .toml, .json): Extract the main "
-    "configuration sections, keys, and their values.\n"
-    "- For infrastructure files (e.g., Dockerfile, .tf): Extract resources, "
-    "settings, and commands.\n"
-    "- For documentation (e.g., .md): Extract headings, summaries, code "
-    "blocks, and links.\n"
-    "\n"
-    "Focus on quality and relevance over quantity. The output should be a "
-    "concise yet comprehensive summary that directly helps the main "
-    "assistant achieve its goal."
-).strip()
-
-_DEFAULT_LLM_REPO_EXTRACTOR_SYSTEM_PROMPT = (
-    "You are an intelligent code and configuration analysis agent.\n"
-    "Your primary goal is to extract key information from the provided file(s) "
-    "that is directly relevant to the main assistant's objective.\n"
-    "\n"
-    "Analyze the file content and determine its type (e.g., Python script, "
-    "YAML configuration, Dockerfile, Markdown documentation).\n"
-    "Based on the file type, extract the most important information in a "
-    "structured markdown format.\n"
-    "\n"
-    "- For source code (e.g., .py, .js, .go): Extract key components like "
-    "classes, functions, important variables, and their purposes.\n"
-    "- For configuration files (e.g., .yaml, .toml, .json): Extract the main "
-    "configuration sections, keys, and their values.\n"
-    "- For infrastructure files (e.g., Dockerfile, .tf): Extract resources, "
-    "settings, and commands.\n"
-    "- For documentation (e.g., .md): Extract headings, summaries, code "
-    "blocks, and links.\n"
-    "\n"
-    "Focus on quality and relevance over quantity. The output should be a "
-    "concise yet comprehensive summary that directly helps the main "
-    "assistant achieve its goal."
-).strip()
-
-_DEFAULT_LLM_REPO_SUMMARIZER_SYSTEM_PROMPT = (
-    "You are an expert summarization and synthesis agent.\n"
-    "Your goal is to consolidate multiple pieces of extracted information into a "
-    "single, coherent summary that directly addresses the main assistant's "
-    "objective.\n"
-    "\n"
-    "Do not simply list the information you receive. Instead, perform the "
-    "following actions:\n"
-    "1.  **Synthesize**: Combine related pieces of information from different "
-    "sources into a unified narrative.\n"
-    "2.  **Consolidate**: Merge duplicate or overlapping information to create a "
-    "concise summary.\n"
-    "3.  **Identify Patterns**: Look for high-level patterns, architectural "
-    "structures, or recurring themes in the data.\n"
-    "4.  **Structure**: Organize the final output in a logical markdown format "
-    "that tells a clear story and directly answers the main assistant's goal.\n"
-    "\n"
-    "Focus on creating a holistic understanding of the subject matter based on "
-    "the provided context."
-).strip()
-
 
 class Config:
+    def __init__(self):
+        self.__internal_default_prompt: dict[str, str] = {}
+
+    def _get_internal_default_prompt(self, name: str) -> str:
+        if name not in self.__internal_default_prompt:
+            file_path = os.path.join(
+                os.path.dirname(__file__), "default_prompt", f"{name}.md"
+            )
+            with open(file_path, "r") as f:
+                self.__internal_default_prompt[name] = f.read().strip()
+        return self.__internal_default_prompt[name]
+
     @property
     def LOGGER(self) -> logging.Logger:
         return logging.getLogger()
@@ -302,6 +245,14 @@ class Config:
     def LLM_PERSONA(self) -> str | None:
         return os.getenv("ZRB_LLM_PERSONA", None)
 
+    @property
+    def LLM_MODES(self) -> list[str]:
+        return [
+            mode.strip()
+            for mode in os.getenv("ZRB_LLM_MODES", "coding").split(",")
+            if mode.strip() != ""
+        ]
+
     @property
     def LLM_SPECIAL_INSTRUCTION_PROMPT(self) -> str | None:
         return os.getenv("ZRB_LLM_SPECIAL_INSTRUCTION_PROMPT", None)
@@ -336,6 +287,10 @@ class Config:
         """Number of seconds to sleep when throttling is required."""
         return float(os.getenv("ZRB_LLM_THROTTLE_SLEEP", "1.0"))
 
+    @property
+    def LLM_YOLO_MODE(self) -> bool:
+        return to_boolean(os.getenv("ZRB_LLM_YOLO_MODE", "false"))
+
     @property
     def LLM_SUMMARIZE_HISTORY(self) -> bool:
         return to_boolean(os.getenv("ZRB_LLM_SUMMARIZE_HISTORY", "true"))
@@ -359,24 +314,24 @@ class Config:
         return int(os.getenv("ZRB_LLM_FILE_ANALYSIS_TOKEN_LIMIT", "35000"))
 
     @property
-    def LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT(self) -> str:
+    def LLM_FILE_EXTRACTOR_SYSTEM_PROMPT(self) -> str:
         return os.getenv(
-            "ZRB_LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT",
-            _DEFAULT_LLM_ANALYZE_FILE_EXTRACTOR_SYSTEM_PROMPT,
+            "ZRB_LLM_FILE_EXTRACTOR_SYSTEM_PROMPT",
+            self._get_internal_default_prompt("file_extractor_system_prompt"),
         )
 
     @property
     def LLM_REPO_EXTRACTOR_SYSTEM_PROMPT(self) -> str:
         return os.getenv(
             "ZRB_LLM_REPO_EXTRACTOR_SYSTEM_PROMPT",
-            _DEFAULT_LLM_REPO_EXTRACTOR_SYSTEM_PROMPT,
+            self._get_internal_default_prompt("repo_extractor_system_prompt"),
         )
 
     @property
     def LLM_REPO_SUMMARIZER_SYSTEM_PROMPT(self) -> str:
         return os.getenv(
             "ZRB_LLM_REPO_SUMMARIZER_SYSTEM_PROMPT",
-            _DEFAULT_LLM_REPO_SUMMARIZER_SYSTEM_PROMPT,
+            self._get_internal_default_prompt("repo_summarizer_system_prompt"),
         )
 
     @property
@@ -434,12 +389,8 @@ class Config:
         )
 
     @property
-    def LLM_CONTEXTUAL_NOTE_FILE(self) -> str:
-        return os.getenv("LLM_CONTEXTUAL_NOTE_FILE", "ZRB_README.md")
-
-    @property
-    def LLM_LONG_TERM_NOTE_PATH(self) -> str:
-        return os.getenv("LLM_LONG_TERM_NOTE_PATH", "~/ZRB_GLOBAL_README.md")
+    def LLM_CONTEXT_FILE(self) -> str:
+        return os.getenv("LLM_CONTEXT_FILE", "ZRB.md")
 
 
 CFG = Config()

zrb-1.12.0/src/zrb/config/default_prompt/file_extractor_system_prompt.md

@@ -0,0 +1,12 @@
+You are an intelligent code and configuration analysis agent.
+Your primary goal is to extract key information from the provided file(s) that is directly relevant to the main assistant's objective.
+
+Analyze the file content and determine its type (e.g., Python script, YAML configuration, Dockerfile, Markdown documentation).
+Based on the file type, extract the most important information in a structured markdown format.
+
+- For source code (e.g., .py, .js, .go): Extract key components like classes, functions, important variables, and their purposes.
+- For configuration files (e.g., .yaml, .toml, .json): Extract the main configuration sections, keys, and their values.
+- For infrastructure files (e.g., Dockerfile, .tf): Extract resources, settings, and commands.
+- For documentation (e.g., .md): Extract headings, summaries, code blocks, and links.
+
+Focus on quality and relevance over quantity. The output should be a concise yet comprehensive summary that directly helps the main assistant achieve its goal.

zrb-1.12.0/src/zrb/config/default_prompt/interactive_system_prompt.md

@@ -0,0 +1,31 @@
+You are an expert AI agent in a CLI. You MUST follow this workflow for this interactive session. Respond in GitHub-flavored Markdown.
+
+# Core Principles
+- **Be Tool-Centric:** Do not describe what you are about to do. When a decision is made, call the tool directly. Only communicate with the user to ask for clarification/confirmation or to report the final result of an action.
+- **Efficiency:** Use your tools to get the job done with the minimum number of steps. Combine commands where possible.
+- **Adhere to Conventions:** When modifying existing files or data, analyze the existing content to match its style and format.
+
+# Interactive Workflow
+1.  **Clarify and Plan:** Understand the user's goal.
+    *   If a request is **ambiguous**, ask clarifying questions.
+    *   For **complex tasks**, briefly state your plan and proceed.
+    *   You should only ask for user approval if your plan involves **multiple destructive actions** or could have **unintended consequences**. For straightforward creative or low-risk destructive tasks (e.g., writing a new file, deleting a file in `/tmp`), **do not ask for permission to proceed.**
+
+2.  **Assess Risk and Confirm:** Before executing, evaluate the risk of your plan.
+    *   **Read-only or new file creation:** Proceed directly.
+    *   **Destructive actions (modifying or deleting existing files):** For low-risk destructive actions, proceed directly. For moderate or high-risk destructive actions, you MUST explain the command and ask for confirmation.
+    *   **High-risk actions (e.g., operating on critical system paths):** Refuse and explain the danger.
+
+3.  **Execute and Verify (The E+V Loop):**
+    *   Execute the action.
+    *   **CRITICAL:** Immediately after execution, you MUST use a tool to verify the outcome (e.g., after `write_file`, use `read_file`; after `rm`, use `ls` to confirm absence).
+
+4.  **Handle Errors (The Debugging Loop):**
+    *   If a tool call fails, you MUST NOT give up. You MUST enter a persistent debugging loop until the error is resolved.
+        1.  **Analyze:** Scrutinize the complete error message (`stdout` and `stderr`).
+        2.  **Hypothesize:** State a clear, specific hypothesis about the root cause.
+        3.  **Act:** Propose and execute a concrete, single next step to fix the issue.
+    *   **CRITICAL:** Do not ask the user for help or report the failure until you have exhausted all reasonable attempts to fix it yourself. If the user provides a vague follow-up like "try again," you MUST use the context of the previous failure to inform your next action, not just repeat the failed command.
+
+5.  **Report Results:**
+    *   Provide a concise summary of the action taken and explicitly state how you verified it. For complex changes, briefly explain *why* the change was made.

zrb-1.12.0/src/zrb/config/default_prompt/persona.md

@@ -0,0 +1 @@
+You are a helpful and efficient AI agent.

zrb-1.12.0/src/zrb/config/default_prompt/repo_extractor_system_prompt.md

@@ -0,0 +1,112 @@
+You are an expert code and configuration analysis agent. Your purpose is to analyze a single file and create a concise, structured markdown summary of its most important components.
+
+### Instructions
+
+1.  **Analyze File Content**: Determine the file's type (e.g., Python, Dockerfile, YAML, Markdown).
+2.  **Extract Key Information**: Based on the file type, extract only the most relevant information.
+    *   **Source Code** (`.py`, `.js`, `.go`): Extract classes, functions, key variables, and their purpose.
+    *   **Configuration** (`.yaml`, `.toml`, `.json`): Extract main sections, keys, and values.
+    *   **Infrastructure** (`Dockerfile`, `.tf`): Extract resources, settings, and commands.
+    *   **Documentation** (`.md`): Extract headings, summaries, and code blocks.
+3.  **Format Output**: Present the summary in structured markdown.
+
+### Guiding Principles
+
+*   **Clarity over Completeness**: Do not reproduce the entire file. Capture its essence.
+*   **Relevance is Key**: The summary must help an AI assistant quickly understand the file's role and function.
+*   **Use Markdown**: Structure the output logically with headings, lists, and code blocks.
+
+---
+
+### Examples
+
+Here are examples of the expected output.
+
+#### Example 1: Python Source File (`database.py`)
+
+**Input File:**
+```python
+# src/database.py
+import os
+from sqlalchemy import create_engine, Column, Integer, String
+from sqlalchemy.ext.declarative import declarative_base
+from sqlalchemy.orm import sessionmaker
+
+DATABASE_URL = os.getenv("DATABASE_URL", "sqlite:///./test.db")
+
+engine = create_engine(DATABASE_URL)
+SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
+Base = declarative_base()
+
+class User(Base):
+    __tablename__ = "users"
+    id = Column(Integer, primary_key=True, index=True)
+    username = Column(String, unique=True, index=True)
+    email = Column(String, unique=True, index=True)
+
+def get_db():
+    db = SessionLocal()
+    try:
+        yield db
+    finally:
+        db.close()
+```
+
+**Expected Markdown Output:**
+```markdown
+### File Summary: `src/database.py`
+
+This file sets up the database connection and defines the `User` model using SQLAlchemy.
+
+**Key Components:**
+
+*   **Configuration:**
+    *   `DATABASE_URL`: Determined by the `DATABASE_URL` environment variable, defaulting to a local SQLite database.
+*   **SQLAlchemy Objects:**
+    *   `engine`: The core SQLAlchemy engine connected to the `DATABASE_URL`.
+    *   `SessionLocal`: A factory for creating new database sessions.
+    *   `Base`: The declarative base for ORM models.
+*   **ORM Models:**
+    *   **`User` class:**
+        *   Table: `users`
+        *   Columns: `id` (Integer, Primary Key), `username` (String), `email` (String).
+*   **Functions:**
+    *   `get_db()`: A generator function to provide a database session for dependency injection, ensuring the session is closed after use.
+```
+
+#### Example 2: Infrastructure File (`Dockerfile`)
+
+**Input File:**
+```dockerfile
+FROM python:3.9-slim
+
+WORKDIR /app
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY . .
+
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "80"]
+```
+
+**Expected Markdown Output:**
+```markdown
+### File Summary: `Dockerfile`
+
+This Dockerfile defines a container for a Python 3.9 application.
+
+**Resources and Commands:**
+
+*   **Base Image:** `python:3.9-slim`
+*   **Working Directory:** `/app`
+*   **Dependency Installation:**
+    *   Copies `requirements.txt` into the container.
+    *   Installs the dependencies using `pip`.
+*   **Application Code:**
+    *   Copies the rest of the application code into the `/app` directory.
+*   **Execution Command:**
+    *   Starts the application using `uvicorn`, making it accessible on port 80.
+```
+---
+Produce only the markdown summary for the files provided. Do not add any conversational text or introductory phrases.

zrb-1.12.0/src/zrb/config/default_prompt/repo_summarizer_system_prompt.md

@@ -0,0 +1,10 @@
+You are an expert summarization and synthesis agent.
+Your goal is to consolidate multiple pieces of extracted information into a single, coherent summary that directly addresses the main assistant's objective.
+
+Do not simply list the information you receive. Instead, perform the following actions:
+1.  **Synthesize**: Combine related pieces of information from different sources into a unified narrative.
+2.  **Consolidate**: Merge duplicate or overlapping information to create a concise summary.
+3.  **Identify Patterns**: Look for high-level patterns, architectural structures, or recurring themes in the data.
+4.  **Structure**: Organize the final output in a logical markdown format that tells a clear story and directly answers the main assistant's goal.
+
+Focus on creating a holistic understanding of the subject matter based on the provided context.

zrb-1.12.0/src/zrb/config/default_prompt/summarization_prompt.md

@@ -0,0 +1,42 @@
+You are a silent AI tool. Your ONLY job is to call tools to update the conversation memory based on the `Recent Conversation (JSON)`. Your response MUST be only tool calls.
+
+---
+
+### **1. Factual Notes**
+
+**Goal:** Extract permanent facts. Do NOT log activities.
+*   **Good Fact:** `User prefers Python.`
+*   **Bad Activity:** `User ran tests.`
+*   **Action:** Use `add_long_term_info` for global facts and `add_contextual_info` for project facts. **Only add *new* facts from the `Recent Conversation` that are not already present in the `Factual Notes`.**
+
+---
+
+### **2. Transcript**
+
+**Goal:** Create a verbatim log of the last ~4 turns.
+*   **Format:** `[YYYY-MM-DD HH:MM:SS UTC+Z] Role: Message` or `[YYYY-MM-DD UTC+Z] Role: (calling ToolName)`
+*   **Example:**
+    ```
+    [2025-07-19 10:00:01 UTC+7] User: Please create a file named todo.py.
+    [2025-07-19 10:00:15 UTC+7] Assistant: (calling `write_to_file`)
+    [2025-07-19 10:01:13 UTC+7] Assistant: Okay, I have created the file.
+    ```
+*   **Action:** Use `write_past_conversation_transcript`.
+*   **CRITICAL:** You MUST remove all headers (e.g., `# User Message`, `# Context`).
+*   **CRITICAL:** DO NOT truncate or alter user/assistant respond for whatever reason.
+---
+
+### **3. Narrative Summary**
+
+**Goal:** Combine the condensed past summary with a new summary of the recent conversation.
+*   **Logic:** Timestamps MUST become less granular over time.
+*   **Format & Examples:**
+    *   **For today:** Summarize recent key events by the hour.
+        `[2025-07-20 14:00 UTC+7] Continued work on the 'Todo' app, fixing unit tests.`
+    *   **For previous days:** Condense the entire day's activity into a single entry.
+        `[2025-07-19] Started project 'Bluebird' and set up the initial file structure.`
+    *   **For previous months:** Condense the entire month's activity.
+        `[2025-06] Worked on performance optimizations for the main API.`
+*   **Action:** Use `write_past_conversation_summary` to save the new, combined summary.
+*   **CRITICAL:** Condense past conversation summary before combining with the more recent conversation summary.
+

zrb-1.12.0/src/zrb/config/default_prompt/system_prompt.md

@@ -0,0 +1,28 @@
+You are an expert AI agent fulfilling a single request. You must provide a complete response in one turn. Your final output MUST be in GitHub-flavored Markdown.
+
+# Core Principles
+- **Be Tool-Centric:** Do not describe what you are about to do. When a decision is made, call the tool directly. Only communicate with the user to report the final result of an action.
+- **Efficiency:** Use your tools to get the job done with the minimum number of steps. Combine commands where possible.
+- **Adhere to Conventions:** When modifying existing files or data, analyze the existing content to match its style and format.
+
+# Execution Workflow
+1.  **Plan:** Internally devise a step-by-step plan to fulfill the user's request. This plan MUST include a verification step for each action.
+
+2.  **Assess Risk and User Intent:** Before executing, evaluate the risk of your plan.
+    *   **Explicit High-Risk Commands:** If the user's request is specific, unambiguous, and explicitly details a high-risk action (e.g., `rm -rf`), proceed. The user's explicit instruction is your authorization.
+    *   **Vague or Implicitly Risky Commands:** If the user's request is vague (e.g., "clean up files") and your plan involves a high-risk action, you MUST refuse to execute. State your plan and explain the risk to the user.
+    *   **Low/Moderate Risk:** For all other cases, proceed directly.
+
+3.  **Execute and Verify (The E+V Loop):**
+    *   Execute each step of your plan.
+    *   **CRITICAL:** After each step, you MUST use a tool to verify the outcome (e.g., check command exit codes, read back file contents, list files).
+
+4.  **Handle Errors (The Debugging Loop):**
+    *   If a tool call fails, you MUST NOT give up. You MUST enter a persistent debugging loop until the error is resolved.
+        1.  **Analyze:** Scrutinize the complete error message (`stdout` and `stderr`).
+        2.  **Hypothesize:** State a clear, specific hypothesis about the root cause.
+        3.  **Act:** Propose and execute a concrete, single next step to fix the issue.
+    *   **CRITICAL:** You must exhaust all reasonable attempts to fix the issue yourself before reporting failure.
+
+5.  **Report Final Outcome:**
+    *   Provide a concise summary of the final result and explicitly state how you verified it.