diff --git a/reviews/auth.h.json b/reviews/auth.h.json
deleted file mode 100644
index 1f7000a..0000000
--- a/reviews/auth.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This source code declares a constant character pointer variable with a value representing an API key.\n\n\n\n// MIT License\n\nconst char* api_key = \"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\";",
-  "review": "**Grade: 3**\n\n---\n\n### Bugs\n- The code does not contain any explicit functional bugs, but exposing an API key in plain text is a security risk.\n\n### Optimizations\n- Instead of hardcoding the API key in the source code, consider fetching it from a secured environment variable or configuration file.\n- Use encryption to further secure sensitive data.\n\n### Good points\n- The code is simple and easily readable.\n- It uses a constant modifier which prevents accidental modification of the API key.\n\n### Summary\nThis code snippet exposes a sensitive API key as a plaintext string, which poses significant security risks depending on its usage. It is essential to manage secrets properly and instead use secure methods for handling API keys and other sensitive information.\n\n### Open source alternatives\n- [dotenv](https://github.com/motdotla/dotenv) for environment variable management in Node.js.\n- [ConfigParser](https://github.com/lorenzwalthert/config) used in Python for managing configuration files.\n- [AWS Secrets Manager](https://aws.amazon.com/secrets-manager/) for secure storage and access of secrets in AWS environments.",
-  "filename": "auth.h",
-  "path": "auth.h",
-  "directory": "",
-  "grade": 3,
-  "size": 347,
-  "line_count": 9
-}
\ No newline at end of file
diff --git a/reviews/auth.h.md b/reviews/auth.h.md
deleted file mode 100644
index d5157ac..0000000
--- a/reviews/auth.h.md
+++ /dev/null
@@ -1,22 +0,0 @@
-**Grade: 3**
-
----
-
-### Bugs
-- The code does not contain any explicit functional bugs, but exposing an API key in plain text is a security risk.
-
-### Optimizations
-- Instead of hardcoding the API key in the source code, consider fetching it from a secured environment variable or configuration file.
-- Use encryption to further secure sensitive data.
-
-### Good points
-- The code is simple and easily readable.
-- It uses a constant modifier which prevents accidental modification of the API key.
-
-### Summary
-This code snippet exposes a sensitive API key as a plaintext string, which poses significant security risks depending on its usage. It is essential to manage secrets properly and instead use secure methods for handling API keys and other sensitive information.
-
-### Open source alternatives
-- [dotenv](https://github.com/motdotla/dotenv) for environment variable management in Node.js.
-- [ConfigParser](https://github.com/lorenzwalthert/config) used in Python for managing configuration files.
-- [AWS Secrets Manager](https://aws.amazon.com/secrets-manager/) for secure storage and access of secrets in AWS environments.
\ No newline at end of file
diff --git a/reviews/cgi-bin/gpt.py.json b/reviews/cgi-bin/gpt.py.json
deleted file mode 100644
index a1354d9..0000000
--- a/reviews/cgi-bin/gpt.py.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".py",
-  "source": "#!/usr/bin/env python3\n\n# Not written by retoor! This is generated boiler plate to give an example!\n\nimport cgi\nimport cgitb\nfrom xmlrpc.client import ServerProxy \nclient = ServerProxy(\"https://api.molodetz.nl/rpc\")\nask_gpt = client.gpt4o_mini\n\ncgitb.enable()\n\nprint(\"Content-Type: text/html\")\nprint()\n\nimport pathlib \n\n\nform = cgi.FieldStorage()\nquestion = form.getvalue(\"question\", \"\")\n\npage_source = pathlib.Path(__file__).parent.joinpath(\"gpt_template.html\").read_text()\n\nif question:\n    try:\n        response = ask_gpt(question)\n    except Exception as e:\n        response = f\"Error: {e}\"\n    page_source = page_source.replace(\"...\", response)\n    page_source = page_source.replace(\"display:none;\",\"\")\n\nprint(page_source)\n",
-  "review": "# 5\n\n## Bugs\n- The code does not handle the case where \"gpt_template.html\" does not exist or is unreadable, which will raise an exception.\n- There's a missing check for the `ask_gpt` function's existence/availability before calling it, potentially causing a runtime error.\n- `cgi` and `cgitb` modules are more suitable for scripting in CGI environments, not for modern web environments; security issues may arise.\n\n## Optimizations\n- Consider using modern frameworks like Flask or Django for handling web requests and templates, instead of the CGI approach.\n- Add error handling for the file reading operation using `pathlib` to catch any potential IO errors.\n- The usage of hardcoded HTML manipulation can be replaced with template engines like Jinja2 for better flexibility and security.\n\n## Good points\n- The use of `pathlib.Path` is a modern, cleaner approach to handle file paths.\n- The script explicitly sets content type and handles basic exception catching.\n\n## Summary\nThe script demonstrates a simple CGI-based web interface to interact with an XML-RPC API that provides GPT4 responses. However, the usage of `cgi` and `cgitb` is outdated for web development; modern frameworks provide better support and security. The script lacks comprehensive error handling and could benefit from enhanced user feedback.\n\n## Open source alternatives\n- **Flask**: A micro web framework written in Python. Ideal for small to medium web applications.\n- **Django**: A high-level Python Web framework that encourages rapid development and clean, pragmatic design.\n- **FastAPI**: A modern, fast (high-performance), web framework for building APIs with Python 3.6+ based on standard Python type hints.",
-  "filename": "gpt.py",
-  "path": "cgi-bin/gpt.py",
-  "directory": "cgi-bin",
-  "grade": 5,
-  "size": 728,
-  "line_count": 33
-}
\ No newline at end of file
diff --git a/reviews/cgi-bin/gpt.py.md b/reviews/cgi-bin/gpt.py.md
deleted file mode 100644
index be9f9ed..0000000
--- a/reviews/cgi-bin/gpt.py.md
+++ /dev/null
@@ -1,23 +0,0 @@
-# 5
-
-## Bugs
-- The code does not handle the case where "gpt_template.html" does not exist or is unreadable, which will raise an exception.
-- There's a missing check for the `ask_gpt` function's existence/availability before calling it, potentially causing a runtime error.
-- `cgi` and `cgitb` modules are more suitable for scripting in CGI environments, not for modern web environments; security issues may arise.
-
-## Optimizations
-- Consider using modern frameworks like Flask or Django for handling web requests and templates, instead of the CGI approach.
-- Add error handling for the file reading operation using `pathlib` to catch any potential IO errors.
-- The usage of hardcoded HTML manipulation can be replaced with template engines like Jinja2 for better flexibility and security.
-
-## Good points
-- The use of `pathlib.Path` is a modern, cleaner approach to handle file paths.
-- The script explicitly sets content type and handles basic exception catching.
-
-## Summary
-The script demonstrates a simple CGI-based web interface to interact with an XML-RPC API that provides GPT4 responses. However, the usage of `cgi` and `cgitb` is outdated for web development; modern frameworks provide better support and security. The script lacks comprehensive error handling and could benefit from enhanced user feedback.
-
-## Open source alternatives
-- **Flask**: A micro web framework written in Python. Ideal for small to medium web applications.
-- **Django**: A high-level Python Web framework that encourages rapid development and clean, pragmatic design.
-- **FastAPI**: A modern, fast (high-performance), web framework for building APIs with Python 3.6+ based on standard Python type hints.
\ No newline at end of file
diff --git a/reviews/cgi-bin/gpt_template.html.json b/reviews/cgi-bin/gpt_template.html.json
deleted file mode 100644
index 6742514..0000000
--- a/reviews/cgi-bin/gpt_template.html.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".html",
-  "source": "<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n    <meta charset=\"UTF-8\">\n    <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n    <title>GPT Example</title>\n    <style>\n        body {\n            font-family: Arial, sans-serif;\n            margin: 0;\n            padding: 0;\n            background-color: #f4f4f9;\n        }\n        .container {\n            max-width: 600px;\n            margin: 50px auto;\n            padding: 20px;\n            background: #fff;\n            border-radius: 8px;\n            box-shadow: 0 4px 10px rgba(0, 0, 0, 0.1);\n        }\n        h1 {\n            color: #333;\n            text-align: center;\n        }\n        textarea, input {\n            width: 100%;\n            padding: 10px;\n            margin: 10px 0;\n            font-size: 16px;\n            border: 1px solid #ddd;\n            border-radius: 4px;\n        }\n        input[type=\"submit\"] {\n            background-color: #4CAF50;\n            color: white;\n            cursor: pointer;\n        }\n        input[type=\"submit\"]:hover {\n            background-color: #45a049;\n        }\n        .response-box {\n            padding: 10px;\n            background: #f9f9f9;\n            border: 1px solid #ddd;\n            border-radius: 4px;\n            min-height: 100px;\n        }\n    </style>\n</head>\n<body>\n    <div class=\"container\">\n        <h1>Ask GPT</h1>\n        <div style=\"display:none;\" class=\"response-box\">\n            <p id=\"response\">...</p>\n        </div>\n        <form action=\"/cgi-bin/gpt.py\" method=\"post\">\n            <textarea name=\"question\" rows=\"4\" placeholder=\"Your prompt..\"></textarea>\n            <input type=\"submit\" value=\"Get Answer\">\n        </form>\n        \n    </div>\n</body>\n</html>",
-  "review": "**7**\n\n### Bugs\n- There are no critical bugs in the HTML structure or CSS.\n\n### Optimizations\n- Use semantic tags like `<main>` or `<section>` for better accessibility and clearer structure instead of using `<div>` everywhere.\n- Consider loading a web font (such as Google Fonts) instead of relying on system fonts for broader compatibility.\n- Replace inline styles with classes and move them into the `<style>` block for consistency and reusability, like the style for the `.response-box`.\n- Consider adding ARIA attributes to make the form more accessible to screen readers.\n  \n### Good points\n- The layout is clean and simple, making it user-friendly.\n- Responsive design with a fluid layout adapts well to different screen sizes due to the maximum-width and margin auto settings.\n- Good use of CSS for visual hierarchy and aesthetics, including color and box-shadow.\n\n### Summary\nThe code presents a simple and clean HTML form styled with CSS, targeting a straightforward user interaction to submit a prompt and receive a response. While it's functional and follows basic HTML/CSS best practices, there is room for improvement in accessibility, semantic HTML, and performance optimizations like font loading.\n\n### Open source alternatives\n- [OpenAI's GPT-3 API Wrapper](https://github.com/openai/gpt-3) - Though more complex, this repository offers a comprehensive implementation using GPT-3.\n- [Hugging Face's Transformers](https://github.com/huggingface/transformers) - Provides interfaces for using transformer models in an application, which includes text generation capabilities similar to GPT.",
-  "filename": "gpt_template.html",
-  "path": "cgi-bin/gpt_template.html",
-  "directory": "cgi-bin",
-  "grade": 7,
-  "size": 1718,
-  "line_count": 64
-}
\ No newline at end of file
diff --git a/reviews/cgi-bin/gpt_template.html.md b/reviews/cgi-bin/gpt_template.html.md
deleted file mode 100644
index 3aed5c9..0000000
--- a/reviews/cgi-bin/gpt_template.html.md
+++ /dev/null
@@ -1,22 +0,0 @@
-**7**
-
-### Bugs
-- There are no critical bugs in the HTML structure or CSS.
-
-### Optimizations
-- Use semantic tags like `<main>` or `<section>` for better accessibility and clearer structure instead of using `<div>` everywhere.
-- Consider loading a web font (such as Google Fonts) instead of relying on system fonts for broader compatibility.
-- Replace inline styles with classes and move them into the `<style>` block for consistency and reusability, like the style for the `.response-box`.
-- Consider adding ARIA attributes to make the form more accessible to screen readers.
-  
-### Good points
-- The layout is clean and simple, making it user-friendly.
-- Responsive design with a fluid layout adapts well to different screen sizes due to the maximum-width and margin auto settings.
-- Good use of CSS for visual hierarchy and aesthetics, including color and box-shadow.
-
-### Summary
-The code presents a simple and clean HTML form styled with CSS, targeting a straightforward user interaction to submit a prompt and receive a response. While it's functional and follows basic HTML/CSS best practices, there is room for improvement in accessibility, semantic HTML, and performance optimizations like font loading.
-
-### Open source alternatives
-- [OpenAI's GPT-3 API Wrapper](https://github.com/openai/gpt-3) - Though more complex, this repository offers a comprehensive implementation using GPT-3.
-- [Hugging Face's Transformers](https://github.com/huggingface/transformers) - Provides interfaces for using transformer models in an application, which includes text generation capabilities similar to GPT.
\ No newline at end of file
diff --git a/reviews/chat.h.json b/reviews/chat.h.json
deleted file mode 100644
index a4e6c0f..0000000
--- a/reviews/chat.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This code defines functionality for creating and managing JSON-based chat prompts \n// using a specific AI model configuration, providing easy integration with message handling \n// and HTTP communication for dynamic applications. \n\n// Non-standard imports: json-c library for handling JSON objects.\n\n\n// MIT License\n//\n// Permission is hereby granted, free of charge, to any person obtaining a copy\n// of this software and associated documentation files (the \"Software\"), to deal\n// in the Software without restriction, including without limitation the rights\n// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n// copies of the Software, and to permit persons to whom the Software is\n// furnished to do so, subject to the following conditions:\n//\n// The above copyright notice and this permission notice shall be included in all\n// copies or substantial portions of the Software.\n//\n// THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n// SOFTWARE.\n\n\n#ifndef CALPACA_PROMPT_H\n#define CALPACA_PROMPT_H\n\n#include <json-c/json.h>\n#include \"messages.h\"\n#include \"http.h\"\n\nchar *prompt_model = \"gpt-4o-mini\";\nint prompt_max_tokens = 100;\ndouble prompt_temperature = 0.5;\n\njson_object *_prompt = NULL;\n\nvoid chat_free() {\n    if (_prompt == NULL)\n        return;\n\n    json_object_put(_prompt);\n    _prompt = NULL;\n}\n\nchar *chat_json(char *role, char *message) {\n    chat_free();\n    message_add(role, message);\n\n    struct json_object *root_object = json_object_new_object();\n    json_object_object_add(root_object, \"model\", json_object_new_string(prompt_model));\n    json_object_object_add(root_object, \"messages\", message_list());\n    json_object_object_add(root_object, \"max_tokens\", json_object_new_int(prompt_max_tokens));\n    json_object_object_add(root_object, \"temperature\", json_object_new_double(prompt_temperature));\n\n    return (char *)json_object_to_json_string_ext(root_object, JSON_C_TO_STRING_PRETTY);\n}\n\n#endif",
-  "review": "**Grade: 6**\n\n### Bugs\n- If `message_add()` or `message_list()` fail or return unexpected results, it might lead to incorrect JSON output without explicit error handling.\n- The return type of `json_object_to_json_string_ext()` is cast to `(char *)`, which may lead to memory management issues if the underlying JSON-C library does not provide a persistently allocated string.\n\n### Optimizations\n- Consider adding error handling for JSON creation and manipulation functions to ensure robustness.\n- Cache the result of `message_list()` if it doesn't change between calls, to improve performance.\n- Ensure the proper handling or casting of the string returned by `json_object_to_json_string_ext()` to prevent possible memory issues.\n- Improve modularity by separating concerns, such as moving JSON object creation into a dedicated function.\n\n### Good points\n- The code uses `json-c` effectively for JSON manipulation, facilitating easy JSON-based data handling.\n- The code is concise and focused on its purpose, with a straightforward structure.\n- Well-defined license and comments improve code understanding and legal clarity.\n\n### Summary\nThe provided code offers a simple yet functional way to create JSON strings that represent chat prompts by leveraging a JSON-handling library. While the implementation is straightforward and concise, adding error handling and optimizing internal function calls could enhance the stability and performance. Also, consider improving memory management based on how the JSON-C library handles string allocations. \n\n### Open source alternatives\n- Rasa (https://rasa.com): An open-source machine learning framework to automate text-and-voice-based assistants.\n- Botpress (https://botpress.com): An open-source conversational AI platform for developing chatbots.\n- Dialogflow (https://dialogflow.cloud.google.com/): Though primarily used as a Google service, the framework allows exporting and using APIs similarly for open-source initiatives.",
-  "filename": "chat.h",
-  "path": "chat.h",
-  "directory": "",
-  "grade": 6,
-  "size": 2399,
-  "line_count": 65
-}
\ No newline at end of file
diff --git a/reviews/chat.h.md b/reviews/chat.h.md
deleted file mode 100644
index 9f09ebb..0000000
--- a/reviews/chat.h.md
+++ /dev/null
@@ -1,24 +0,0 @@
-**Grade: 6**
-
-### Bugs
-- If `message_add()` or `message_list()` fail or return unexpected results, it might lead to incorrect JSON output without explicit error handling.
-- The return type of `json_object_to_json_string_ext()` is cast to `(char *)`, which may lead to memory management issues if the underlying JSON-C library does not provide a persistently allocated string.
-
-### Optimizations
-- Consider adding error handling for JSON creation and manipulation functions to ensure robustness.
-- Cache the result of `message_list()` if it doesn't change between calls, to improve performance.
-- Ensure the proper handling or casting of the string returned by `json_object_to_json_string_ext()` to prevent possible memory issues.
-- Improve modularity by separating concerns, such as moving JSON object creation into a dedicated function.
-
-### Good points
-- The code uses `json-c` effectively for JSON manipulation, facilitating easy JSON-based data handling.
-- The code is concise and focused on its purpose, with a straightforward structure.
-- Well-defined license and comments improve code understanding and legal clarity.
-
-### Summary
-The provided code offers a simple yet functional way to create JSON strings that represent chat prompts by leveraging a JSON-handling library. While the implementation is straightforward and concise, adding error handling and optimizing internal function calls could enhance the stability and performance. Also, consider improving memory management based on how the JSON-C library handles string allocations. 
-
-### Open source alternatives
-- Rasa (https://rasa.com): An open-source machine learning framework to automate text-and-voice-based assistants.
-- Botpress (https://botpress.com): An open-source conversational AI platform for developing chatbots.
-- Dialogflow (https://dialogflow.cloud.google.com/): Though primarily used as a Google service, the framework allows exporting and using APIs similarly for open-source initiatives.
\ No newline at end of file
diff --git a/reviews/http.h.json b/reviews/http.h.json
deleted file mode 100644
index 31cb576..0000000
--- a/reviews/http.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// The source code provides functionality for making HTTP POST and GET requests over SSL/TLS using OpenSSL. It includes initialization and cleanup of the OpenSSL library, creation of SSL context, socket creation and connection, and sending requests with handling responses. Furthermore, it interfaces with JSON and handles authentication using an external \"auth.h\" file.\n\n// Includes: \"auth.h\", <json-c/json.h>\n\n// MIT License\n\n#ifndef CALPACA_HTTP_H\n#define CALPACA_HTTP_H\n\n#include <stdio.h>\n#include <string.h>\n#include <stdlib.h>\n#include <unistd.h>\n#include <arpa/inet.h>\n#include <netdb.h>\n#include <openssl/ssl.h>\n#include <openssl/err.h>\n#include \"auth.h\"\n#include <json-c/json.h>\n\nvoid init_openssl() {\n    SSL_load_error_strings();\n    OpenSSL_add_ssl_algorithms();\n}\n\nvoid cleanup_openssl() {\n    EVP_cleanup();\n}\n\nSSL_CTX *create_context() {\n    const SSL_METHOD *method = TLS_method();\n    SSL_CTX *ctx = SSL_CTX_new(method);\n    SSL_CTX_load_verify_locations(ctx, \"/etc/ssl/certs/ca-certificates.crt\", NULL);\n\n    return ctx;\n}\n\nSSL_CTX *create_context2() {\n    const SSL_METHOD *method = TLS_client_method();\n    SSL_CTX *ctx = SSL_CTX_new(method);\n    if (!ctx) {\n        perror(\"Unable to create SSL context\");\n        ERR_print_errors_fp(stderr);\n        exit(EXIT_FAILURE);\n    }\n\n    return ctx;\n}\n\nint create_socket(const char *hostname, int port) {\n    struct hostent *host;\n    struct sockaddr_in addr;\n\n    host = gethostbyname(hostname);\n    if (!host) {\n        perror(\"Unable to resolve host\");\n        exit(EXIT_FAILURE);\n    }\n\n    int sock = socket(AF_INET, SOCK_STREAM, 0);\n    if (sock < 0) {\n        perror(\"Unable to create socket\");\n        exit(EXIT_FAILURE);\n    }\n\n    addr.sin_family = AF_INET;\n    addr.sin_port = htons(port);\n    addr.sin_addr.s_addr = *(long *)(host->h_addr);\n\n    if (connect(sock, (struct sockaddr *)&addr, sizeof(addr)) != 0) {\n        perror(\"Unable to connect to host\");\n        close(sock);\n        exit(EXIT_FAILURE);\n    }\n\n    return sock;\n}\n\nchar *http_post(const char *hostname, char *url, char *data) {\n    init_openssl();\n    int port = 443;\n    SSL_CTX *ctx = create_context();\n    int sock = create_socket(hostname, port);\n    SSL *ssl = SSL_new(ctx);\n    SSL_set_connect_state(ssl);\n    SSL_set_tlsext_host_name(ssl, hostname);\n    SSL_set_fd(ssl, sock);\n\n    int buffer_size = 4096;\n    char *buffer = malloc(buffer_size);\n    \n    if (SSL_connect(ssl) <= 0) {\n        ERR_print_errors_fp(stderr);\n    } else {\n        size_t len = strlen(data);\n        char *request = malloc(len + 4096);\n        sprintf(request,\n                \"POST %s HTTP/1.1\\r\\n\"\n                \"Content-Length: %ld\\r\\n\"\n                \"Content-Type: application/json\\r\\n\"\n                \"Host: api.openai.com\\r\\n\"\n                \"Authorization: Bearer %s\\r\\n\"\n                \"Connection: close\\r\\n\\r\\n%s\",\n                url, len, api_key, data);\n        \n        SSL_write(ssl, request, strlen(request));\n        free(request);\n\n        int bytes;\n        int bytes_total = 0;\n        while ((bytes = SSL_read(ssl, buffer + bytes_total, buffer_size - 1)) > 0) {\n            bytes_total += bytes;\n            buffer = realloc(buffer, bytes_total + buffer_size);\n            buffer[bytes_total] = '\\0';\n        }\n    }\n\n    SSL_free(ssl);\n    close(sock);\n    SSL_CTX_free(ctx);\n    cleanup_openssl();\n    \n    return buffer;\n}\n\nchar *http_get(const char *hostname, char *url) {\n    init_openssl();\n    int port = 443;\n    SSL_CTX *ctx = create_context();\n    int sock = create_socket(hostname, port);\n    SSL *ssl = SSL_new(ctx);\n    SSL_set_connect_state(ssl);\n    SSL_set_tlsext_host_name(ssl, hostname);\n    SSL_set_fd(ssl, sock);\n\n    int buffer_size = 4096;\n    char *buffer = malloc(buffer_size);\n\n    if (SSL_connect(ssl) <= 0) {\n        ERR_print_errors_fp(stderr);\n    } else {\n        char request[buffer_size];\n        sprintf(request,\n                \"GET %s HTTP/1.1\\r\\n\"\n                \"Host: api.openai.com\\r\\n\"\n                \"Authorization: Bearer %s\\r\\n\"\n                \"Connection: close\\r\\n\\r\\n\",\n                url, api_key);\n\n        SSL_write(ssl, request, strlen(request));\n\n        int bytes;\n        int bytes_total = 0;\n        while ((bytes = SSL_read(ssl, buffer + bytes_total, buffer_size - 1)) > 0) {\n            bytes_total += bytes;\n            buffer = realloc(buffer, bytes_total + buffer_size);\n            buffer[bytes_total] = '\\0';\n        }\n    }\n\n    SSL_free(ssl);\n    close(sock);\n    SSL_CTX_free(ctx);\n    cleanup_openssl();\n\n    return buffer;\n}\n#endif",
-  "review": "# 6\n\n## Bugs\n- `create_context` and `create_context2` functions are redundant; however, only `create_context2` performs error checking, which is crucial.\n- Use of `gethostbyname` is deprecated; it should be replaced with more modern alternatives like `getaddrinfo`.\n- `api_key` is used within `http_post` and `http_get` functions without being defined in the code.\n- Risk of memory leak when reallocating the buffer if `realloc` fails, the original buffer is not freed, and the system would have reduced memory.\n- `close(sock)` does not check for errors.\n\n## Optimizations\n- Consolidate `create_context` and `create_context2` into a single function to avoid redundancy and potential maintenance issues.\n- Use `getaddrinfo` instead of `gethostbyname` for better compatibility and thread safety.\n- Include proper error handling if `realloc` fails to make efficient memory management.\n- Specify size when using `malloc` or `realloc` for buffer to enhance readability and maintenance.\n- Use `snprintf` instead of `sprintf` to avoid buffer overflow vulnerabilities.\n\n## Good points\n- Proper initialization and cleanup of OpenSSL.\n- Effective use of OpenSSL APIs to set up TLS/SSL connections.\n- Clear function separation for handling different tasks like initializing SSL, creating sockets, and managing HTTP requests.\n- Usage of dynamic memory management for buffer handling shows good understanding.\n\n## Summary\nThe code demonstrates an understanding of OpenSSL API for creating secure HTTP connections over SSL/TLS, and effectively segregates the key functionalities. However, there are a few significant issues that need addressing, including handling deprecated functions, managing possible memory leaks, and ensuring error checking across all parts of the code. The use of global `api_key` is problematic since its source isn't evident within the code snippet. Memory handling and duplication of functions are areas where optimizations could significantly enhance the robustness and efficiency of the code.\n\n## Open source alternatives\n- **Libcurl**: A widely used library that offers a simple and consistent API for making HTTP requests, supporting a variety of protocols and features including SSL/TLS.\n- **Boost.Beast**: This C++ library is part of the Boost collection, providing an HTTP and WebSocket client and server built on top of Boost.Asio. It offers SSL/TLS support via OpenSSL.\n- **HttpClient in Poco**: The POCO C++ Libraries offer components that simplify HTTP/S communications, including robust support for SSL/TLS.",
-  "filename": "http.h",
-  "path": "http.h",
-  "directory": "",
-  "grade": 6,
-  "size": 4593,
-  "line_count": 170
-}
\ No newline at end of file
diff --git a/reviews/http.h.md b/reviews/http.h.md
deleted file mode 100644
index 2203f1b..0000000
--- a/reviews/http.h.md
+++ /dev/null
@@ -1,29 +0,0 @@
-# 6
-
-## Bugs
-- `create_context` and `create_context2` functions are redundant; however, only `create_context2` performs error checking, which is crucial.
-- Use of `gethostbyname` is deprecated; it should be replaced with more modern alternatives like `getaddrinfo`.
-- `api_key` is used within `http_post` and `http_get` functions without being defined in the code.
-- Risk of memory leak when reallocating the buffer if `realloc` fails, the original buffer is not freed, and the system would have reduced memory.
-- `close(sock)` does not check for errors.
-
-## Optimizations
-- Consolidate `create_context` and `create_context2` into a single function to avoid redundancy and potential maintenance issues.
-- Use `getaddrinfo` instead of `gethostbyname` for better compatibility and thread safety.
-- Include proper error handling if `realloc` fails to make efficient memory management.
-- Specify size when using `malloc` or `realloc` for buffer to enhance readability and maintenance.
-- Use `snprintf` instead of `sprintf` to avoid buffer overflow vulnerabilities.
-
-## Good points
-- Proper initialization and cleanup of OpenSSL.
-- Effective use of OpenSSL APIs to set up TLS/SSL connections.
-- Clear function separation for handling different tasks like initializing SSL, creating sockets, and managing HTTP requests.
-- Usage of dynamic memory management for buffer handling shows good understanding.
-
-## Summary
-The code demonstrates an understanding of OpenSSL API for creating secure HTTP connections over SSL/TLS, and effectively segregates the key functionalities. However, there are a few significant issues that need addressing, including handling deprecated functions, managing possible memory leaks, and ensuring error checking across all parts of the code. The use of global `api_key` is problematic since its source isn't evident within the code snippet. Memory handling and duplication of functions are areas where optimizations could significantly enhance the robustness and efficiency of the code.
-
-## Open source alternatives
-- **Libcurl**: A widely used library that offers a simple and consistent API for making HTTP requests, supporting a variety of protocols and features including SSL/TLS.
-- **Boost.Beast**: This C++ library is part of the Boost collection, providing an HTTP and WebSocket client and server built on top of Boost.Asio. It offers SSL/TLS support via OpenSSL.
-- **HttpClient in Poco**: The POCO C++ Libraries offer components that simplify HTTP/S communications, including robust support for SSL/TLS.
\ No newline at end of file
diff --git a/reviews/line.h.json b/reviews/line.h.json
deleted file mode 100644
index 49209e7..0000000
--- a/reviews/line.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This source code provides command-line input functionalities with autocomplete and history features using readline library functionalities. It allows users to complete commands and manage input history.\n\n// External includes: \n// - <readline/readline.h>\n// - <readline/history.h>\n\n// MIT License: Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction.\n\n#include <readline/readline.h>\n#include <readline/history.h>\n\n#define HISTORY_FILE \"~/.calpaca_history\"\n\nbool line_initialized = false;\n\nchar* line_command_generator(const char* text, int state) {\n    static int list_index, len;\n    const char* commands[] = {\"help\", \"exit\", \"list\", \"review\", \"refactor\", \"obfuscate\", NULL};\n\n    if (!state) {\n        list_index = 0;\n        len = strlen(text);\n    }\n\n    while (commands[list_index]) {\n        const char* command = commands[list_index++];\n        if (strncmp(command, text, len) == 0) {\n            return strdup(command);\n        }\n    }\n\n    return NULL;\n}\n\nchar** line_command_completion(const char* text, int start, int end) {\n    rl_attempted_completion_over = 1;\n    return rl_completion_matches(text, line_command_generator);\n}\n\nvoid line_init() {\n    if (!line_initialized) {\n        rl_attempted_completion_function = line_command_completion;\n        line_initialized = true;\n        read_history(HISTORY_FILE);\n    }\n}\n\nchar* line_read(char* prefix) {\n    char* data = readline(prefix);\n    if (!(data && *data)) {\n        return NULL;\n    }\n    return data;\n}\n\nvoid line_add_history(char* data) {\n    read_history(HISTORY_FILE);\n    add_history(data);\n    write_history(HISTORY_FILE);\n}",
-  "review": "8\n\n## Bugs\n- No bugs identified; the code appears to function as intended.\n\n## Optimizations\n- Use `realpath` to ensure `HISTORY_FILE` resolves correctly and consistently, mitigating issues with tilde expansion.\n- Ensure `line_initialized` is thread-safe if accessed from multiple threads.\n- Consider freeing duplicated strings to avoid memory leaks.\n\n## Good points\n- Efficient handling of command completions using a static array of known commands.\n- Proper use of the readline library functions for history and completion features.\n- The code is concise and focused on specific functionalities of input handling.\n\n## Summary\nThe provided C code uses the readline library to facilitate command-line input with autocomplete and history functionality. It is well-written, adhering to proper memory management practices, though some minor optimizations could be considered. The implementation is straightforward, offering essential features like command completion and history saving for a better user experience in command-line applications.\n\n## Open source alternatives\n- [GNU Readline](https://tiswww.case.edu/php/chet/readline/rltop.html): A library that already provides similar functionalities and more, widely used and maintained.\n- [linenoise](https://github.com/antirez/linenoise): A small self-contained alternative to readline that supports history and completion.\n- [libedit](http://thrysoee.dk/editline/): Another more lightweight alternative that offers a BSD-licensed readline replacement.",
-  "filename": "line.h",
-  "path": "line.h",
-  "directory": "",
-  "grade": 8,
-  "size": 1777,
-  "line_count": 62
-}
\ No newline at end of file
diff --git a/reviews/line.h.md b/reviews/line.h.md
deleted file mode 100644
index 9439cd6..0000000
--- a/reviews/line.h.md
+++ /dev/null
@@ -1,22 +0,0 @@
-8
-
-## Bugs
-- No bugs identified; the code appears to function as intended.
-
-## Optimizations
-- Use `realpath` to ensure `HISTORY_FILE` resolves correctly and consistently, mitigating issues with tilde expansion.
-- Ensure `line_initialized` is thread-safe if accessed from multiple threads.
-- Consider freeing duplicated strings to avoid memory leaks.
-
-## Good points
-- Efficient handling of command completions using a static array of known commands.
-- Proper use of the readline library functions for history and completion features.
-- The code is concise and focused on specific functionalities of input handling.
-
-## Summary
-The provided C code uses the readline library to facilitate command-line input with autocomplete and history functionality. It is well-written, adhering to proper memory management practices, though some minor optimizations could be considered. The implementation is straightforward, offering essential features like command completion and history saving for a better user experience in command-line applications.
-
-## Open source alternatives
-- [GNU Readline](https://tiswww.case.edu/php/chet/readline/rltop.html): A library that already provides similar functionalities and more, widely used and maintained.
-- [linenoise](https://github.com/antirez/linenoise): A small self-contained alternative to readline that supports history and completion.
-- [libedit](http://thrysoee.dk/editline/): Another more lightweight alternative that offers a BSD-licensed readline replacement.
\ No newline at end of file
diff --git a/reviews/main.c.json b/reviews/main.c.json
deleted file mode 100644
index 55476dd..0000000
--- a/reviews/main.c.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".c",
-  "source": "// Written by retoor@molodetz.nl\n\n// This source code initializes a command-line application that uses OpenAI for chat interactions, handles user inputs, and can start a simple HTTP server with CGI support. The code allows command execution, markdown parsing, and OpenAI chat integration.\n\n// External imports used in this code:\n// - openai.h\n// - markdown.h\n// - plugin.h\n// - line.h\n\n// MIT License\n//\n// Permission is hereby granted, free of charge, to any person obtaining a copy\n// of this software and associated documentation files (the \"Software\"), to deal\n// in the Software without restriction, including without limitation the rights\n// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n// copies of the Software, and to permit persons to whom the Software is\n// furnished to do so, subject to the following conditions:\n//\n// The above copyright notice and this permission notice shall be included in\n// all copies or substantial portions of the Software.\n//\n// THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN\n// THE SOFTWARE.\n\n#include \"openai.h\"\n#include \"markdown.h\"\n#include \"plugin.h\"\n#include \"line.h\"\n#include <locale.h>\n#include <stdio.h>\n#include <stdlib.h>\n#include <string.h>\n#include <unistd.h>\n\nchar *get_prompt_from_args(int c, char **argv) {\n    char *prompt = malloc(1024 * 1024 + 1);\n    prompt[0] = 0;\n    for (int i = 1; i < c; i++) {\n        if (argv[i][0] == '-')\n            break;\n        strncat(prompt, argv[i], 1024 * 1024);\n        if (i < c - 1) {\n            strncat(prompt, \" \", 1024 * 1024);\n        } else {\n            strncat(prompt, \".\", 1024 * 1024);\n        }\n    }\n    if (!*prompt) {\n        free(prompt);\n        return NULL;\n    }\n    return prompt;\n}\n\nbool try_prompt(int argc, char *argv[]) {\n    char *prompt = get_prompt_from_args(argc, argv);\n    if (prompt != NULL) {\n        char *response = openai_chat(\"user\", prompt);\n        parse_markdown_to_ansi(response);\n        printf(\"\\n\");\n        free(response);\n        free(prompt);\n        return true;\n    }\n    return false;\n}\n\nvoid help();\nvoid render(char *);\nvoid serve() {\n    render(\"Starting server. *Put executables in a dir named cgi-bin and they will behave as web pages.*\");\n    int res = system(\"python3 -m http.server --cgi\");\n    (void)res;\n}\n\nvoid render(char *content) {\n    parse_markdown_to_ansi(content);\n    printf(\"\\n\\n\");\n}\n\nvoid repl() {\n    line_init();\n    setbuf(stdout, NULL);\n    char *line;\n    char *previous_line = NULL;\n    while ((line = line_read(\"> \"))) {\n        if (!line || !*line) {\n            line = previous_line;\n        }\n        if (!line || !*line)\n            continue;\n        previous_line = line;\n        if (line[0] == '!') {\n            plugin_run(line + 1);\n            continue;\n        }\n        if (!strncmp(line, \"exit\", 4)) {\n            exit(0);\n        }\n        if (!strncmp(line, \"help\", 4)) {\n            help();\n            continue;\n        }\n        if (!strncmp(line, \"serve\", 5)) {\n            serve();\n        }\n        if (!strncmp(line, \"spar \", 5)) {\n            char *response = line + 5;\n            while (true) {\n                render(response);\n                sleep(2);\n                response = openai_chat(\"user\", response);\n            }\n        }\n        if (!strncmp(line, \"ls\", 2) || !strncmp(line, \"list\", 4)) {\n            int offset = 2;\n            if (!strncmp(line, \"list\", 4)) {\n                offset = 4;\n            }\n            char *command = (char *)malloc(strlen(line) + 42);\n            command[0] = 0;\n            strcpy(command, \"ls \");\n            strcat(command, line + offset);\n            int res = system(command);\n            (void)res;\n            free(command);\n            continue;\n        }\n\n        line_add_history(line);\n        char *response = openai_chat(\"user\", line);\n        render(response);\n        free(response);\n    }\n}\n\nvoid help() {\n    char help_text[1024 * 1024] = {0};\n    char *template = \"# Help\\n\"\n                     \"Written by retoor@molodetz.nl.\\n\\n\"\n                     \"## Features\\n\"\n                     \" - navigate through history using `arrows`.\\n\"\n                     \" - navigate through history with **recursive search** using `ctrl+r`.\\n\"\n                     \" - **inception with python** for *incoming* and *outgoing* content.\\n\"\n                     \" - markdown and **syntax highlighting**.\\n\"\n                     \" - **execute python commands** with prefix `!`\\n\"\n                     \" - list files of the current work directory using `ls`.\\n\"\n                     \" - type `serve` to start a web server with directory listing. Easy for network transfers.\\n\\n\"\n                     \"## Configuration\\n\"\n                     \" - model temperature is %f.\\n\"\n                     \" - model name is %s.\\n\"\n                     \" - max tokens is %d.\\n\\n\"\n                     \"## In development\\n\"\n                     \" - **google search** and actions with those results.\\n\"\n                     \" - **reminders**.\\n\"\n                     \" - predefined **templates** for **reviewing** / **refactoring** so you can personalize.\\n\";\n    sprintf(help_text, template, prompt_temperature, prompt_model, prompt_max_tokens);\n    render(help_text);\n}\n\nvoid openai_include(char *path) {\n    FILE *file = fopen(path, \"r\");\n    if (file == NULL) {\n        return;\n    }\n    fseek(file, 0, SEEK_END);\n    long size = ftell(file);\n    fseek(file, 0, SEEK_SET);\n\n    char *buffer = (char *)malloc(size);\n    size_t read = fread(buffer, 1, size, file);\n    if (read == 0) {\n        return;\n    }\n\n    fclose(file);\n    buffer[read] = 0;\n    openai_system(buffer);\n\n    free(buffer);\n}\n\nvoid init() {\n    line_init();\n    const char *locale = setlocale(LC_ALL, NULL);\n    char payload[4096] = {0};\n    sprintf(payload, \"User locale is %s. User lang is %s.\\n\"\n                    \"You are Retoor. Use a lot of markdown in response.\\n\"\n                    \"Be confident and short in answers.\\n\"\n                    \"You divide things by zero if you have to.\",\n            locale, locale);\n\n    printf(\"%s\", \"Loading...\");\n    openai_system(payload);\n    openai_include(\"context.txt\");\n    printf(\"%s\", \"\\rLoaded! Type help for features.\\n\");\n}\n\nint main(int argc, char *argv[]) {\n    init();\n    if (try_prompt(argc, argv))\n        return 0;\n\n    repl();\n    return 0;\n}",
-  "review": "**Grade: 6**\n\n### Bugs\n- Potential dereference of a null pointer `line` in `repl()` after the `previous_line` assignment and check.\n- In `get_prompt_from_args()`, using `strncat` without precise bounds checking could lead to buffer overflow.\n- In `openai_include()`, `fread` does not handle the case where reading fewer bytes than `size`, which could leave the buffer uninitialized.\n- Missing check for the return value of `malloc`, leading to potential null pointer dereference.\n\n### Optimizations\n- Use `snprintf` or `asprintf` to dynamically manage buffer sizes instead of fixed large allocation sizes in functions like `get_prompt_from_args()`.\n- Replace the system command calls with native functions to avoid potential security risks.\n- Consider sharing string operations' results like in `command` formation in `repl()` using a more efficient concatenation process.\n- Implement a graceful shutdown mechanism in `repl()` when terminating the application.\n\n### Good Points\n- The code uses a modular approach, with functions tailored to specific tasks.\n- Good use of external libraries to offload complex tasks such as markdown parsing.\n- The code allows flexibility by using command-line arguments and REPL for user interaction.\n- The use of clear and descriptive comments makes the codebase easier to understand and maintain.\n\n### Summary\nThe code provides a feature-rich command-line application leveraging OpenAI and other libraries for various functions. While functional and well-commented, it contains some bugs and potential security risks from system command execution. Optimizations can further streamline and secure the logic, especially regarding memory management and error handling. \n\n### Open source alternatives\n- [Rasa](https://rasa.com/) provides open-source tools for building custom conversational applications.\n- [ChatterBot](https://github.com/gunthercox/ChatterBot) is an open-source conversational dialog engine used for building chatbots.\n- [Botpress](https://botpress.com/) is another open-source conversational AI platform suitable for similar use cases.",
-  "filename": "main.c",
-  "path": "main.c",
-  "directory": "",
-  "grade": 6,
-  "size": 6758,
-  "line_count": 212
-}
\ No newline at end of file
diff --git a/reviews/main.c.md b/reviews/main.c.md
deleted file mode 100644
index f24398f..0000000
--- a/reviews/main.c.md
+++ /dev/null
@@ -1,27 +0,0 @@
-**Grade: 6**
-
-### Bugs
-- Potential dereference of a null pointer `line` in `repl()` after the `previous_line` assignment and check.
-- In `get_prompt_from_args()`, using `strncat` without precise bounds checking could lead to buffer overflow.
-- In `openai_include()`, `fread` does not handle the case where reading fewer bytes than `size`, which could leave the buffer uninitialized.
-- Missing check for the return value of `malloc`, leading to potential null pointer dereference.
-
-### Optimizations
-- Use `snprintf` or `asprintf` to dynamically manage buffer sizes instead of fixed large allocation sizes in functions like `get_prompt_from_args()`.
-- Replace the system command calls with native functions to avoid potential security risks.
-- Consider sharing string operations' results like in `command` formation in `repl()` using a more efficient concatenation process.
-- Implement a graceful shutdown mechanism in `repl()` when terminating the application.
-
-### Good Points
-- The code uses a modular approach, with functions tailored to specific tasks.
-- Good use of external libraries to offload complex tasks such as markdown parsing.
-- The code allows flexibility by using command-line arguments and REPL for user interaction.
-- The use of clear and descriptive comments makes the codebase easier to understand and maintain.
-
-### Summary
-The code provides a feature-rich command-line application leveraging OpenAI and other libraries for various functions. While functional and well-commented, it contains some bugs and potential security risks from system command execution. Optimizations can further streamline and secure the logic, especially regarding memory management and error handling. 
-
-### Open source alternatives
-- [Rasa](https://rasa.com/) provides open-source tools for building custom conversational applications.
-- [ChatterBot](https://github.com/gunthercox/ChatterBot) is an open-source conversational dialog engine used for building chatbots.
-- [Botpress](https://botpress.com/) is another open-source conversational AI platform suitable for similar use cases.
\ No newline at end of file
diff --git a/reviews/markdown.h.json b/reviews/markdown.h.json
deleted file mode 100644
index 0a5675e..0000000
--- a/reviews/markdown.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This program provides functionality to highlight keywords in source code with ANSI color formatting and to convert Markdown syntax into ANSI-colored text output.\n\n// Uses standard C libraries: <stdio.h>, <string.h>. Also utilizes ANSI escape codes for text formatting.\n\n// MIT License:\n// Permission is hereby granted, free of charge, to any person obtaining a copy\n// of this software and associated documentation files (the \"Software\"), to deal\n// in the Software without restriction, including without limitation the rights\n// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n// copies of the Software, and to permit persons to whom the Software is\n// furnished to do so, subject to the following conditions:\n//\n// The above copyright notice and this permission notice shall be included in all\n// copies or substantial portions of the Software.\n//\n// THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n// SOFTWARE.\n\n#include <stdio.h>\n#include <string.h>\n#include <stdbool.h>\n\n#define RESET \"\\033[0m\"\n#define BOLD \"\\033[1m\"\n#define ITALIC \"\\033[3m\"\n#define FG_YELLOW \"\\033[33m\"\n#define FG_BLUE \"\\033[34m\"\n#define FG_CYAN \"\\033[36m\"\n\nint is_keyword(const char *word) {\n    const char *keywords[] = {\"int\", \"float\", \"double\", \"char\", \"void\", \"if\", \"else\", \"while\", \"for\", \"return\", \"struct\", \"printf\"};\n    for (size_t i = 0; i < sizeof(keywords) / sizeof(keywords[0]); i++) {\n        if (strcmp(word, keywords[i]) == 0) {\n            return 1;\n        }\n    }\n    return 0;\n}\n\nvoid highlight_code(const char *code) {\n    const char *ptr = code;\n    char buffer[256];\n    size_t index = 0;\n\n    while (*ptr) {\n        if ((*ptr >= 'a' && *ptr <= 'z') || (*ptr >= 'A' && *ptr <= 'Z') || (*ptr == '_')) {\n            while ((*ptr >= 'a' && *ptr <= 'z') || (*ptr >= 'A' && *ptr <= 'Z') || (*ptr >= '0' && *ptr <= '9') || (*ptr == '_')) {\n                buffer[index++] = *ptr++;\n            }\n            buffer[index] = '\\0';\n\n            if (is_keyword(buffer)) {\n                printf(FG_BLUE \"%s\" RESET, buffer);\n            } else {\n                printf(\"%s\", buffer);\n            }\n            index = 0;\n        } else if (*ptr >= '0' && *ptr <= '9') {\n            while (*ptr >= '0' && *ptr <= '9') {\n                buffer[index++] = *ptr++;\n            }\n            buffer[index] = '\\0';\n            printf(FG_CYAN \"%s\" RESET, buffer);\n            index = 0;\n        } else {\n            putchar(*ptr);\n            ptr++;\n        }\n    }\n}\n\nvoid parse_markdown_to_ansi(const char *markdown) {\n    const char *ptr = markdown;\n    bool inside_code = false;\n\n    while (*ptr) {\n        if (*ptr == '`') {\n            inside_code = !inside_code;\n            if (inside_code) {\n                printf(FG_YELLOW);\n            } else {\n                printf(RESET);\n            }\n            ptr++;\n            continue;\n        }\n\n        if (inside_code) {\n            char code_buffer[256];\n            size_t index = 0;\n\n            while (*ptr && *ptr != '`') {\n                code_buffer[index++] = *ptr++;\n            }\n            code_buffer[index] = '\\0';\n            highlight_code(code_buffer);\n        } else {\n            if (strncmp(ptr, \"**\", 2) == 0) {\n                printf(BOLD);\n                ptr += 2;\n                while (*ptr && strncmp(ptr, \"**\", 2) != 0) putchar(*ptr++);\n                if (*ptr == '*' && *(ptr + 1) == '*') ptr += 2;\n                printf(RESET);\n            }\n            else if (*ptr == '*' && (ptr == markdown || *(ptr - 1) != '*')) {\n                printf(ITALIC);\n                ptr++;\n                while (*ptr && *ptr != '*') putchar(*ptr++);\n                if (*ptr == '*') ptr++;\n                printf(RESET);\n            }\n            else if (strncmp(ptr, \"### \", 4) == 0) {\n                printf(BOLD FG_YELLOW);\n                ptr += 4;\n                while (*ptr && *ptr != '\\n') putchar(*ptr++);\n                printf(RESET \"\\n\");\n            } else if (strncmp(ptr, \"## \", 3) == 0) {\n                printf(BOLD FG_YELLOW);\n                ptr += 3;\n                while (*ptr && *ptr != '\\n') putchar(*ptr++);\n                printf(RESET \"\\n\");\n            } else if (strncmp(ptr, \"# \", 2) == 0) {\n                printf(BOLD FG_YELLOW);\n                ptr += 2;\n                while (*ptr && *ptr != '\\n') putchar(*ptr++);\n                printf(RESET \"\\n\");\n            } else {\n                putchar(*ptr);\n                ptr++;\n            }\n        }\n    }\n}",
-  "review": "**7**\n\n### Bugs\n- Potential buffer overflow in `highlight_code` and `parse_markdown_to_ansi` functions.\n- Use of hard-coded buffer size which may not be enough for large input.\n\n### Optimizations\n- Use dynamic strings or vector-like data structures to handle larger inputs safely.\n- Optimize the `is_keyword` function by using a hash set or another efficient data structure for keyword lookup.\n- Refactor common code patterns (like the `while` loops for parsing) into reusable helper functions to reduce redundancy.\n  \n### Good Points\n- Clear use of ANSI escape codes for formatting which enhances code readability.\n- Proper division between handling code syntax and markdown syntax.\n- Inclusion of both code highlighting and markdown parsing in a single, cohesive codebase.\n- Follows good practices regarding licensing with the inclusion of the MIT License.\n\n### Summary\nThe code efficiently uses ANSI escape codes to highlight keywords in source code and parse markdown into ANSI-colored text. The main concerns are related to the safety of buffer usage, as it is prone to overflow when handling long inputs. The keyword detection could be optimized for speed, and the code could be made more reusable with refactoring. Overall, the program serves its purpose well but would benefit from improvements in robustness and efficiency.\n\n### Open source alternatives\n- **Pygments**: A generic syntax highlighter that supports many languages and formats.\n- **highlight.js**: A JavaScript library with similar functionality that works in browsers.\n- **Rouge**: A Ruby-based syntax highlighter compatible with over 100 languages.",
-  "filename": "markdown.h",
-  "path": "markdown.h",
-  "directory": "",
-  "grade": 7,
-  "size": 4933,
-  "line_count": 140
-}
\ No newline at end of file
diff --git a/reviews/markdown.h.md b/reviews/markdown.h.md
deleted file mode 100644
index de61c11..0000000
--- a/reviews/markdown.h.md
+++ /dev/null
@@ -1,24 +0,0 @@
-**7**
-
-### Bugs
-- Potential buffer overflow in `highlight_code` and `parse_markdown_to_ansi` functions.
-- Use of hard-coded buffer size which may not be enough for large input.
-
-### Optimizations
-- Use dynamic strings or vector-like data structures to handle larger inputs safely.
-- Optimize the `is_keyword` function by using a hash set or another efficient data structure for keyword lookup.
-- Refactor common code patterns (like the `while` loops for parsing) into reusable helper functions to reduce redundancy.
-  
-### Good Points
-- Clear use of ANSI escape codes for formatting which enhances code readability.
-- Proper division between handling code syntax and markdown syntax.
-- Inclusion of both code highlighting and markdown parsing in a single, cohesive codebase.
-- Follows good practices regarding licensing with the inclusion of the MIT License.
-
-### Summary
-The code efficiently uses ANSI escape codes to highlight keywords in source code and parse markdown into ANSI-colored text. The main concerns are related to the safety of buffer usage, as it is prone to overflow when handling long inputs. The keyword detection could be optimized for speed, and the code could be made more reusable with refactoring. Overall, the program serves its purpose well but would benefit from improvements in robustness and efficiency.
-
-### Open source alternatives
-- **Pygments**: A generic syntax highlighter that supports many languages and formats.
-- **highlight.js**: A JavaScript library with similar functionality that works in browsers.
-- **Rouge**: A Ruby-based syntax highlighter compatible with over 100 languages.
\ No newline at end of file
diff --git a/reviews/messages.h.json b/reviews/messages.h.json
deleted file mode 100644
index ff9073a..0000000
--- a/reviews/messages.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This code manages a collection of messages using JSON objects. It provides functions to retrieve all messages as a JSON array, add a new message with a specified role and content, and free the allocated resources.\n\n// Includes external library <json-c/json.h> for JSON manipulation\n\n// MIT License\n// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:\n// The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.\n// THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.\n\n#ifndef CALPACA_MESSAGES_H\n#define CALPACA_MESSAGES_H\n#include \"json-c/json.h\"\n\nstruct json_object *_message_array = NULL;\n\nstruct json_object *message_list() {\n    if (_message_array == NULL) {\n        _message_array = json_object_new_array();\n    }\n    return _message_array;\n}\n\nstruct json_object *message_add(char *role, char *content) {\n    struct json_object *messages = message_list();\n    struct json_object *message = json_object_new_object();\n    json_object_object_add(message, \"role\", json_object_new_string(role));\n    json_object_object_add(message, \"content\", json_object_new_string(content));\n    json_object_array_add(messages, message);\n    return message;\n}\n\nchar *message_json() {\n    return (char *)json_object_to_json_string_ext(message_list(), JSON_C_TO_STRING_PRETTY);\n}\n\nvoid message_free() {\n    if (_message_array != NULL) {\n        json_object_put(_message_array);\n        _message_array = NULL;\n    }\n}\n#endif",
-  "review": "8\n\n## Bugs\n- None identified in the basic functionality provided.\n\n## Optimizations\n- Convert the global variable `_message_array` into a static, thread-local variable if multithreading is applied in the project to avoid race conditions.\n- Consider passing constant strings as `const char *role` and `const char *content` to the `message_add` function to ensure safety and potential optimization.\n- Use explicit memory management strategies if `<json-c/json.h>` library supports them for better handling of resources beyond `json_object_put`.\n\n## Good points\n- Proper use of the `json-c` library for JSON manipulation.\n- The code is clean, with meaningful function names that align with their purpose.\n- Implements proper memory deallocation with `message_free` function ensuring no memory leaks.\n- Includes detailed comments explaining the logic and the licensing, contributing to the code's transparency and its legal use.\n\n## Summary\nThis code demonstrates effective use of the `json-c` library to handle JSON data structures, providing a simple API for managing a collection of messages. It is well-commented, maintains good practices for JSON operations, and ensures memory management by freeing resources with `message_free`. Small improvements regarding the mutability of string arguments and thread safety in multithreaded environments may enhance its robustness further. The absence of identified bugs is a testament to the meticulous nature of the implementation.\n\n## Open source alternatives\n- **Jansson**: A C library for encoding, decoding, and manipulating JSON data.\n- **CJSON**: Ultralightweight JSON parser in ANSI C.",
-  "filename": "messages.h",
-  "path": "messages.h",
-  "directory": "",
-  "grade": 8,
-  "size": 2304,
-  "line_count": 44
-}
\ No newline at end of file
diff --git a/reviews/messages.h.md b/reviews/messages.h.md
deleted file mode 100644
index a8da588..0000000
--- a/reviews/messages.h.md
+++ /dev/null
@@ -1,22 +0,0 @@
-8
-
-## Bugs
-- None identified in the basic functionality provided.
-
-## Optimizations
-- Convert the global variable `_message_array` into a static, thread-local variable if multithreading is applied in the project to avoid race conditions.
-- Consider passing constant strings as `const char *role` and `const char *content` to the `message_add` function to ensure safety and potential optimization.
-- Use explicit memory management strategies if `<json-c/json.h>` library supports them for better handling of resources beyond `json_object_put`.
-
-## Good points
-- Proper use of the `json-c` library for JSON manipulation.
-- The code is clean, with meaningful function names that align with their purpose.
-- Implements proper memory deallocation with `message_free` function ensuring no memory leaks.
-- Includes detailed comments explaining the logic and the licensing, contributing to the code's transparency and its legal use.
-
-## Summary
-This code demonstrates effective use of the `json-c` library to handle JSON data structures, providing a simple API for managing a collection of messages. It is well-commented, maintains good practices for JSON operations, and ensures memory management by freeing resources with `message_free`. Small improvements regarding the mutability of string arguments and thread safety in multithreaded environments may enhance its robustness further. The absence of identified bugs is a testament to the meticulous nature of the implementation.
-
-## Open source alternatives
-- **Jansson**: A C library for encoding, decoding, and manipulating JSON data.
-- **CJSON**: Ultralightweight JSON parser in ANSI C.
\ No newline at end of file
diff --git a/reviews/openai.h.json b/reviews/openai.h.json
deleted file mode 100644
index 4537149..0000000
--- a/reviews/openai.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This code provides functions to interact with OpenAI's APIs. It includes functionalities for fetching available models, system interactions, and engaging in chat-based conversations using the OpenAI API.\n\n\n// Imports the \"http\" library for handling HTTP requests and the \"chat\" library for JSON handling related to chat content.\n\n\n// MIT License: Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files, to deal in the Software without restriction, including the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.\n\n\n#ifndef CALPACA_OPENAI_H\n#define CALPACA_OPENAI_H\n#include \"http.h\"\n#include \"chat.h\"\n#include <string.h>\n#include <stdbool.h>\n\nchar *openai_get_models() {\n    const char *hostname = \"api.openai.com\";\n    char *url = \"/v1/models\";\n    return http_get(hostname, url);\n}\n\nbool openai_system(char *content) {\n    const char *hostname = \"api.openai.com\";\n    char *url = \"/v1/chat/completions\";\n    char *data = chat_json(\"system\", content);\n    char *result = http_post(hostname, url, data);\n    bool is_done = result != NULL;\n    \n    free(result);\n    return is_done;\n}\n\nchar *openai_chat(char *role, char *content) {\n    const char *hostname = \"api.openai.com\";\n    char *url = \"/v1/chat/completions\";\n    char *data = chat_json(role, content);\n    char *result = http_post(hostname, url, data);\n    char *body = strstr(result, \"\\r\\n\\r\\n\") + 4;\n    body = strstr(body, \"\\r\\n\");\n    body = strstr(body, \"\\r\\n\");\n    *(body - 5) = 0;\n    struct json_object *parsed_json = json_tokener_parse(body);\n    if (!parsed_json) {\n        fprintf(stderr, \"Failed to parse JSON.\\n\");\n        return NULL;\n    }\n\n    struct json_object *choices_array;\n    if (!json_object_object_get_ex(parsed_json, \"choices\", &choices_array)) {\n        fprintf(stderr, \"Failed to get 'choices' array.\\n\");\n        json_object_put(parsed_json);\n        return NULL;\n    }\n\n    struct json_object *first_choice = json_object_array_get_idx(choices_array, 0);\n    if (!first_choice) {\n        fprintf(stderr, \"Failed to get the first element of 'choices'.\\n\");\n        json_object_put(parsed_json);\n        return NULL;\n    }\n\n    struct json_object *message_object;\n    if (!json_object_object_get_ex(first_choice, \"message\", &message_object)) {\n        fprintf(stderr, \"Failed to get 'message' object.\\n\");\n        json_object_put(parsed_json);\n        return NULL;\n    }\n\n    char *content_str = (char *)json_object_get_string(json_object_object_get(message_object, \"content\"));\n    message_add(\"assistant\", content_str);\n    free(data);\n    free(result);\n    char *final_result = strdup(content_str);\n   \n    json_object_put(parsed_json);\n\n    return final_result;\n}\n\n#endif",
-  "review": "# 7\n\n## Bugs\n- **Memory Management**: Potential memory leaks due to missing `free()` calls for dynamically allocated strings (`data`, `result`) in `openai_chat`.\n- **Buffer Overflows**: Pointer arithmetic in `openai_chat` might lead to undefined behavior if there's an unexpected response format.\n- **Invalid Access**: In `openai_chat`, the manipulation to null-terminate the response string (`*(body - 5) = 0;`) depends on assumptions that might not always hold, potentially causing segmentation faults.\n\n## Optimizations\n- **Memory Management**: Ensure all dynamically allocated memory is properly freed to prevent memory leaks.\n- **Error Handling**: Improve handling for a wider range of possible errors in network interactions and JSON parsing.\n- **Reuse Connections**: Use persistent connections for HTTP requests if supported by the library to reduce latency.\n- **String Handling**: Use safer string manipulation methods to avoid buffer overflows and segmentation faults.\n- **Modularization**: Consider splitting functionalities into smaller functions to improve code readability and maintainability.\n\n## Good points\n- **Library Usage**: Correctly utilizes external libraries to manage HTTP requests and JSON parsing.\n- **Code Organization**: The separation of functionalities into different functions is clear and well-structured.\n- **License Clarity**: The inclusion of the MIT license ensures clarity on usage permissions.\n\n## Summary\nOverall, this code provides a straightforward implementation for interacting with OpenAI's APIs using basic HTTP request and JSON handling mechanisms. However, there are several areas, especially concerning memory management and error handling, that require improvement for increased reliability and efficiency. With optimizations in these areas, the code could be robust enough for production environments. \n\n## Open source alternatives\n- **Langchain**: A framework for developing applications with language models which can interface with OpenAI.\n- **GPT-3 Python Client**: A Python package for interacting with OpenAI's APIs.\n- **openai-cpp**: A C++ wrapper for OpenAI's API which may cover similar functionalities with additional abstractions for ease of use.",
-  "filename": "openai.h",
-  "path": "openai.h",
-  "directory": "",
-  "grade": 7,
-  "size": 3517,
-  "line_count": 83
-}
\ No newline at end of file
diff --git a/reviews/openai.h.md b/reviews/openai.h.md
deleted file mode 100644
index 87266ed..0000000
--- a/reviews/openai.h.md
+++ /dev/null
@@ -1,26 +0,0 @@
-# 7
-
-## Bugs
-- **Memory Management**: Potential memory leaks due to missing `free()` calls for dynamically allocated strings (`data`, `result`) in `openai_chat`.
-- **Buffer Overflows**: Pointer arithmetic in `openai_chat` might lead to undefined behavior if there's an unexpected response format.
-- **Invalid Access**: In `openai_chat`, the manipulation to null-terminate the response string (`*(body - 5) = 0;`) depends on assumptions that might not always hold, potentially causing segmentation faults.
-
-## Optimizations
-- **Memory Management**: Ensure all dynamically allocated memory is properly freed to prevent memory leaks.
-- **Error Handling**: Improve handling for a wider range of possible errors in network interactions and JSON parsing.
-- **Reuse Connections**: Use persistent connections for HTTP requests if supported by the library to reduce latency.
-- **String Handling**: Use safer string manipulation methods to avoid buffer overflows and segmentation faults.
-- **Modularization**: Consider splitting functionalities into smaller functions to improve code readability and maintainability.
-
-## Good points
-- **Library Usage**: Correctly utilizes external libraries to manage HTTP requests and JSON parsing.
-- **Code Organization**: The separation of functionalities into different functions is clear and well-structured.
-- **License Clarity**: The inclusion of the MIT license ensures clarity on usage permissions.
-
-## Summary
-Overall, this code provides a straightforward implementation for interacting with OpenAI's APIs using basic HTTP request and JSON handling mechanisms. However, there are several areas, especially concerning memory management and error handling, that require improvement for increased reliability and efficiency. With optimizations in these areas, the code could be robust enough for production environments. 
-
-## Open source alternatives
-- **Langchain**: A framework for developing applications with language models which can interface with OpenAI.
-- **GPT-3 Python Client**: A Python package for interacting with OpenAI's APIs.
-- **openai-cpp**: A C++ wrapper for OpenAI's API which may cover similar functionalities with additional abstractions for ease of use.
\ No newline at end of file
diff --git a/reviews/plugin.h.json b/reviews/plugin.h.json
deleted file mode 100644
index f7288d5..0000000
--- a/reviews/plugin.h.json
+++ /dev/null
@@ -1,11 +0,0 @@
-{
-  "extension": ".h",
-  "source": "// Written by retoor@molodetz.nl\n\n// This source code initializes a Python interpreter within a plugin, executes a provided Python script with some basic imports, and finalizes the Python environment when done.\n\n// This code does not use any non-standard imports or includes aside from Python.h and structmember.h which are part of Python's C API.\n\n// MIT License\n\n\n#include <python3.14/Python.h>\n#include <python3.14/structmember.h>\n#include <stdbool.h>\n#include <stdio.h>\n#include <stdlib.h>\n#include <string.h>\n\nbool plugin_initialized = false;\n\nbool plugin_construct() {\n    if (plugin_initialized)\n        return true;\n\n    Py_Initialize();\n\n    if (!Py_IsInitialized()) {\n        fprintf(stderr, \"Failed to initialize Python interpreter\\n\");\n        return plugin_initialized;\n    }\n    plugin_initialized = true;\n    return plugin_initialized;\n}\n\nvoid plugin_run(char *src) {\n    plugin_construct();\n    const char *basics = \n        \"import sys\\n\"\n        \"import os\\n\"\n        \"import math\\n\"\n        \"import pathlib\\n\"\n        \"import subprocess\\n\"\n        \"import time\\n\"\n        \"from datetime import datetime\\n\"\n        \"%s\";\n    size_t length = strlen(basics) + strlen(src);\n    char *script = (char *)malloc(length + 1);\n    sprintf(script, basics, src);\n    script[length] = '\\0';\n    PyRun_SimpleString(script);\n    free(script);\n}\n\nvoid plugin_destruct() {\n    if (plugin_initialized)\n        Py_Finalize();\n}",
-  "review": "## 5\n\n### Bugs\n- There is no check to see if `PyRun_SimpleString` runs successfully.\n- The allocation with `malloc` is not checked for a `NULL` return, which may lead to a segmentation fault if memory allocation fails.\n  \n### Optimizations\n- Check for successful memory allocation after calling `malloc`.\n- `Py_Initialize()` should ideally be paired with `PyErr_Occurred()` to detect initialization errors.\n- Consider using modern C string formatting functions like `snprintf` to prevent buffer overflow.\n- Ensure the script is valid Python code by checking with `PyErr_Occurred()` after `PyRun_SimpleString`.\n- Use `Py_FinalizeEx()` instead of `Py_Finalize()` when available, as it provides better error handling.\n\n### Good points\n- The plugin checks whether the Python interpreter is already initialized, preventing redundant initialization.\n- Use of `fprintf` for error reporting ensures error messages are printed to the standard error stream.\n- Dynamic construction of the Python script with basic imports is done neatly.\n\n### Summary\nThe code provides a basic mechanism to run a Python interpreter from a C plugin, dynamically constructing the script with necessary imports. Although functionally correct, there are areas requiring error handling and optimization, such as checking memory allocations and Python interpreter errors. These improvements enhance stability and reliability.\n\n### Open source alternatives\n- **CPython** serves as a standard environment to execute Python code embedded in C programs.\n- **Boost.Python** provides a framework for interfacing C++ and Python.\n- **SWIG** simplifies the task of interfacing different languages, including C/C++ with Python.",
-  "filename": "plugin.h",
-  "path": "plugin.h",
-  "directory": "",
-  "grade": 5,
-  "size": 1427,
-  "line_count": 55
-}
\ No newline at end of file
diff --git a/reviews/plugin.h.md b/reviews/plugin.h.md
deleted file mode 100644
index 860a83f..0000000
--- a/reviews/plugin.h.md
+++ /dev/null
@@ -1,25 +0,0 @@
-## 5
-
-### Bugs
-- There is no check to see if `PyRun_SimpleString` runs successfully.
-- The allocation with `malloc` is not checked for a `NULL` return, which may lead to a segmentation fault if memory allocation fails.
-  
-### Optimizations
-- Check for successful memory allocation after calling `malloc`.
-- `Py_Initialize()` should ideally be paired with `PyErr_Occurred()` to detect initialization errors.
-- Consider using modern C string formatting functions like `snprintf` to prevent buffer overflow.
-- Ensure the script is valid Python code by checking with `PyErr_Occurred()` after `PyRun_SimpleString`.
-- Use `Py_FinalizeEx()` instead of `Py_Finalize()` when available, as it provides better error handling.
-
-### Good points
-- The plugin checks whether the Python interpreter is already initialized, preventing redundant initialization.
-- Use of `fprintf` for error reporting ensures error messages are printed to the standard error stream.
-- Dynamic construction of the Python script with basic imports is done neatly.
-
-### Summary
-The code provides a basic mechanism to run a Python interpreter from a C plugin, dynamically constructing the script with necessary imports. Although functionally correct, there are areas requiring error handling and optimization, such as checking memory allocations and Python interpreter errors. These improvements enhance stability and reliability.
-
-### Open source alternatives
-- **CPython** serves as a standard environment to execute Python code embedded in C programs.
-- **Boost.Python** provides a framework for interfacing C++ and Python.
-- **SWIG** simplifies the task of interfacing different languages, including C/C++ with Python.
\ No newline at end of file