Repository navigation
Agent API_p9
Previous: API_p8.md
Depends on: lazygui/config/constants.py, lazygui/services/models.py
Imported by: lazygui/panels/graph_panel.py, lazygui/widgets/__init__.py, tests/test_lazygui_graph_widget.py
-
GraphNodeItem.__init__(method)lazygui/widgets/graph_view.py:143def __init__(self, node, radius, color, pixmap, on_selected, on_context_menu, label_visible, parent) -
GraphNodeItem.node_data(method)lazygui/widgets/graph_view.py:186def node_data(self)-- Return the immutable node data associated with this item. -
GraphNodeItem.paint(method)lazygui/widgets/graph_view.py:190def paint(self, painter, option, widget) -
GraphNodeItem.mousePressEvent(method)lazygui/widgets/graph_view.py:248def mousePressEvent(self, event) -
GraphNodeItem.mouseReleaseEvent(method)lazygui/widgets/graph_view.py:259def mouseReleaseEvent(self, event) -
GraphNodeItem.hoverEnterEvent(method)lazygui/widgets/graph_view.py:266def hoverEnterEvent(self, event) -
GraphNodeItem.hoverLeaveEvent(method)lazygui/widgets/graph_view.py:273def hoverLeaveEvent(self, event) -
GraphNodeItem.itemChange(method)lazygui/widgets/graph_view.py:279def itemChange(self, change, value) -
GraphEdgeItem.__init__(method)lazygui/widgets/graph_view.py:288def __init__(self, edge, source_item, target_item, parent) -
GraphEdgeItem.update_position(method)lazygui/widgets/graph_view.py:324def update_position(self)-- Recalculate the line endpoints from source and target positions. -
GraphEdgeItem.edge_data(method)lazygui/widgets/graph_view.py:329def edge_data(self)-- Return the immutable edge data. -
GraphScene.__init__(method)lazygui/widgets/graph_view.py:340def __init__(self, constants, parent) -
GraphScene.set_topology(method)lazygui/widgets/graph_view.py:349def set_topology(self, topology)-- Replace the current graph with a new topology using force layout. -
GraphScene.step_physics(method)lazygui/widgets/graph_view.py:419def step_physics(self)-- Run one iteration of force-directed layout. -
GraphScene.selected_node_id(method)lazygui/widgets/graph_view.py:491def selected_node_id(self)-- Return the identifier of the first selected node, or None. -
GraphView.__init__(method)lazygui/widgets/graph_view.py:509def __init__(self, constants, parent) -
GraphView.set_theme_colors(method)lazygui/widgets/graph_view.py:527def set_theme_colors(self, bg_color, edge_color)-- Update graph background and default edge colour from theme tokens. -
GraphView.set_topology(method)lazygui/widgets/graph_view.py:535def set_topology(self, topology)-- Render a new graph topology. -
GraphView.wheelEvent(method)lazygui/widgets/graph_view.py:551def wheelEvent(self, event) -
GraphView.selected_node_id(method)lazygui/widgets/graph_view.py:561def selected_node_id(self)-- Return the identifier of the currently selected node. -
GraphView.fit_to_content(method)lazygui/widgets/graph_view.py:565def fit_to_content(self)-- Zoom to fit all graph content. -
GraphView.scene_handle(method)lazygui/widgets/graph_view.py:570def scene_handle(self)-- Return the underlying graph scene.
Depends on: lazygui/services/backend.py
Imported by: lazygui/widgets/__init__.py, lazygui/windows/main_window.py
-
StatusBadge.__init__(method)lazygui/widgets/status_badge.py:21def __init__(self, parent)-- Create the badge in the disconnected state. -
StatusBadge.set_status(method)lazygui/widgets/status_badge.py:27def set_status(self, status)-- Update label text and the object name driving QSS colour.
Depends on: lazygui/config/constants.py
Imported by: lazygui/panels/terminal_panel.py, lazygui/widgets/__init__.py
-
TerminalView.__init__(method)lazygui/widgets/terminal_view.py:38def __init__(self, constants, parent)-- Configure the widget for log-style append-only behaviour. -
TerminalView.append_output(method)lazygui/widgets/terminal_view.py:54def append_output(self, text)-- Appendtextafter stripping ANSI control codes and control chars. -
TerminalView.keyPressEvent(method)lazygui/widgets/terminal_view.py:75def keyPressEvent(self, event)-- Forward keystrokes to listeners instead of mutating the buffer.
Depends on: lazygui/config/constants.py, lazygui/widgets/command_palette_list.py
Imported by: lazygui/windows/__init__.py, lazygui/windows/main_window.py
-
CommandPaletteWindow.__init__(method)lazygui/windows/command_palette_window.py:18def __init__(self, constants, actions, parent)-- Build the search input plus the result list. -
CommandPaletteWindow.set_actions(method)lazygui/windows/command_palette_window.py:43def set_actions(self, actions)-- Replace the action set on the underlying list. -
CommandPaletteWindow.keyPressEvent(method)lazygui/windows/command_palette_window.py:48def keyPressEvent(self, event)-- Intercept Enter and Esc, otherwise delegate to the search box. -
CommandPaletteWindow.showEvent(method)lazygui/windows/command_palette_window.py:71def showEvent(self, event)-- Move keyboard focus into the search box on every open.
Depends on: lazygui/config/c2_credentials.py, lazygui/config/constants.py, lazygui/config/paths.py, lazygui/config/settings.py, lazygui/services/models.py, lazygui/services/teamserver_backend.py
Imported by: lazygui/app.py, lazygui/windows/__init__.py
-
ConnectDialog.__init__(method)lazygui/windows/connect_dialog.py:53def __init__(self, constants, settings, paths, parent)-- Build the form and pre-fill from persisted settings + credentials file. -
ConnectDialog.request(method)lazygui/windows/connect_dialog.py:105def request(self)-- Translate current widget state into a :class:ConnectionRequest. -
ConnectDialog.persist_choice(method)lazygui/windows/connect_dialog.py:120def persist_choice(self)-- Save the picked values back into :class:AppSettings.
Depends on: lazygui/config/constants.py, lazygui/config/settings.py, lazygui/panels/registry.py, lazygui/services/backend.py, lazygui/services/event_log.py, lazygui/services/models.py, lazygui/theme/manager.py, lazygui/theme/tokens.py, lazygui/widgets/beacon_command_modal.py, lazygui/widgets/command_palette_list.py, lazygui/widgets/status_badge.py, lazygui/windows/command_palette_window.py
Imported by: lazygui/app.py, lazygui/windows/__init__.py
-
MainWindow.__init__(method)lazygui/windows/main_window.py:40def __init__(self, constants, settings, theme_manager, backend, event_log, parent)-- Wire the dependencies and build the full layout. -
MainWindow.closeEvent(method)lazygui/windows/main_window.py:99def closeEvent(self, event)-- Persist geometry, layout and theme before closing. -
MainWindow.window_requests_connect(method)lazygui/windows/main_window.py:376def window_requests_connect(self)-- Hook overridden by :class:Applicationto swap backend at runtime. -
MainWindow.status_badge(method)lazygui/windows/main_window.py:393def status_badge(self)-- Expose the badge so the application can update it on backend swaps. -
MainWindow.panels(method)lazygui/windows/main_window.py:398def panels(self)-- Expose the panel registry for backend swaps.
-
extract_ips_from_arp(function)lazyown-docker/hostdiscover.sh:21 -
extract_listening_ips_from_netstat(function)lazyown-docker/hostdiscover.sh:26
-
usage(function)lazyown-docker/mkdocker.sh:23-- Help message -
log(function)lazyown-docker/mkdocker.sh:36-- Log function -
check_docker(function)lazyown-docker/mkdocker.sh:41-- Check if Docker is running -
check_file(function)lazyown-docker/mkdocker.sh:49-- Check if file exists -
container_exists(function)lazyown-docker/mkdocker.sh:58-- Check if container exists -
image_exists(function)lazyown-docker/mkdocker.sh:63-- Check if image exists -
get_ports(function)lazyown-docker/mkdocker.sh:68-- Get ports from payload.json -
build_image(function)lazyown-docker/mkdocker.sh:93-- Build Docker image -
validate_payload(function)lazyown-docker/mkdocker.sh:109-- Validate payload.json exists -
run_container(function)lazyown-docker/mkdocker.sh:117-- Run container -
stop_container(function)lazyown-docker/mkdocker.sh:187-- Stop container -
clean_container_and_image(function)lazyown-docker/mkdocker.sh:197-- Clean container and image
Depends on: cli/aliases.py, cli/auto_crypto.py, cli/autosuggest.py, cli/chain_mode.py, cli/cli_enhancements.py, cli/command_chain.py, cli/engagement_hooks.py, cli/exploration.py, cli/fuzzy_picker.py, cli/graph_advisor.py, cli/headless.py, cli/lazynmap_post.py, cli/ops_commands.py, cli/palette.py, cli/palette_command.py, cli/protips.py, cli/reactive_hints.py, cli/registry.py, cli/scope_guard.py, cli/session_brief.py, cli/splash.py, cli/status_bar.py, cli/tips_engine.py, cli/toast_bus.py, core/autonomous.py, core/config.py, core/console.py, core/credential_vault.py, core/defaults.py, core/hardening.py, core/logging.py, core/param_registry.py, modules/cli_auth.py, modules/command_executor.py, modules/event_bus.py, modules/event_consumers.py, modules/lazy_loop.py, modules/llm_factory.py, modules/logging_config.py, modules/metrics.py, modules/payload_factory.py, modules/session_cleanup.py, modules/state_manager.py, modules/unified_bridge.py, skills/unified_orchestrator.py, utils.py
Imported by: core/command_bridge.py, discord_c2.py, lazyc2.py, poc_tui/app.py, scripts/devtools/command_audit.py, slack_c2_bot.py, telegram_c2.py, tests/test_c2_agent.py, tests/test_scope_guard_integration.py
-
_PayloadSettableProxy.__init__(method)lazyown.py:244def __init__(self, params)-- Bind the proxy to a liveparamsdictionary. -
LazyOwnShell.__init__(method)lazyown.py:372def __init__(self)-- Initializer for the LazyOwnShell class. -
LazyOwnShell.log_command(method)lazyown.py:785def log_command(self, cmd_name, cmd_args, start_time, end_time, duration_ms)-- Logs the command execution details to a CSV file. -
LazyOwnShell.default(method)lazyown.py:848def default(self, line)-- Handles undefined commands, including aliases. -
LazyOwnShell.scripts(method)lazyown.py:877def scripts(self)-- Auto-discovered list of runnable script names. -
LazyOwnShell.refresh_prompt(method)lazyown.py:896def refresh_prompt(self)-- Recompute the Neon Box prompt from the live payload. -
LazyOwnShell.do_set(method)lazyown.py:907def do_set(self, line)-- Set a parameter — the unifiedset/assignsurface. -
LazyOwnShell.logcsv(method)lazyown.py:1288def logcsv(self, line, start_time, end_time, duration_ms)-- Forward a command line to :meth:log_commandfor CSV persistence. -
LazyOwnShell.cmd(method)lazyown.py:1329def cmd(self, line)-- Internal function to execute commands. -
LazyOwnShell.onecmd_plus_hooks(method)lazyown.py:1421def onecmd_plus_hooks(self, statement, add_to_history, raise_keyboard_interrupt, orig_rl_history_length)-- Dispatch a command, expanding payload placeholders in custom aliases. -
LazyOwnShell.one_cmd(method)lazyown.py:1785def one_cmd(self, command)-- Internal function to execute commands. -
LazyOwnShell.emptyline(method)lazyown.py:1831def emptyline(self)-- Handle the case where the user enters an empty line. -
LazyOwnShell.load_user_commands(method)lazyown.py:1856def load_user_commands(self)-- Carga los comandos personalizados desde user_commands.json -
LazyOwnShell.save_user_command(method)lazyown.py:1868def save_user_command(self, alias, command)-- Guarda un nuevo comando en user_commands.json -
LazyOwnShell.list_files_in_directory(method)lazyown.py:1880def list_files_in_directory(self, directory)-- Lista todos los archivos en un directorio dado. -
LazyOwnShell.register_tool_commands(method)lazyown.py:1886def register_tool_commands(self)-- Register every activetools/*.toolas ado_<toolname>command. -
make_wrapper(function)lazyown.py:1945def make_wrapper(cmd_template, tname, default_target) -
tool_wrapper(function)lazyown.py:1946def tool_wrapper(arg) -
wrapper(function)lazyown.py:2047def wrapper(arg) -
LazyOwnShell.load_plugins(method)lazyown.py:2070def load_plugins(self)-- Load every Lua plugin from the 'plugins/' directory. -
LazyOwnShell.load_yaml_plugins(method)lazyown.py:2100def load_yaml_plugins(self)-- Loads all YAML plugins from the 'lazyaddons/' directory. -
LazyOwnShell.register_yaml_plugin(method)lazyown.py:2123def register_yaml_plugin(self, plugin_data)-- Register a YAML addon as a shell command. -
wrapper_yaml(function)lazyown.py:2163def wrapper_yaml(arg) -
LazyOwnShell.register_all_adversary_commands(method)lazyown.py:2269def register_all_adversary_commands(self) -
cmd_wrapper(function)lazyown.py:2291def cmd_wrapper(_) -
LazyOwnShell.display_toastr(method)lazyown.py:2298def display_toastr(self, message, type)-- Display the unified toastr panel from the shared toast bus. -
LazyOwnShell.completedefault(method)lazyown.py:2311def completedefault(self, text, line, begidx, endidx)-- Fall through to the payload-aware completer for unhandled commands. -
LazyOwnShell.preloop(method)lazyown.py:2336def preloop(self)-- Print a session-start pro tip and handle first-run setup. -
LazyOwnShell.postparsing_precmd(method)lazyown.py:2480def postparsing_precmd(self, statement)-- Gate unauthenticated commands — anonymous operators can only runregister,login,logout,whoami... -
LazyOwnShell.postloop(method)lazyown.py:2508def postloop(self)-- Handle operations to perform after exiting the command loop. -
LazyOwnShell.complete_phase(method)lazyown.py:2541def complete_phase(self, text, line, begidx, endidx)-- Tab-complete phase names. -
LazyOwnShell.complete_l00t(method)lazyown.py:2546def complete_l00t(self, text, line, begidx, endidx)-- Tab-complete l00t subcommands. -
LazyOwnShell.complete_loot(method)lazyown.py:2552def complete_loot(self, text, line, begidx, endidx)-- Tab-complete loot subcommands (delegates to l00t). -
LazyOwnShell.complete_assign(method)lazyown.py:2557def complete_assign(self, text, line, begidx, endidx)-- Tab-complete the parameter name from the live payload keys. -
LazyOwnShell.complete_scope(method)lazyown.py:2574def complete_scope(self, text, line, begidx, endidx)-- Tab-complete the scope subcommands. -
LazyOwnShell.complete_palette(method)lazyown.py:2625def complete_palette(self, text, line, begidx, endidx)-- Tab-complete the palette command using the live command index. -
LazyOwnShell.run_lazysearch(method)lazyown.py:2642def run_lazysearch(self)-- Runs the internal modulemodules/lazysearch.py. -
LazyOwnShell.run_lazysearch_gui(method)lazyown.py:2659def run_lazysearch_gui(self)-- Run the internal module located atmodules/LazyOwnExplorer.py. -
LazyOwnShell.run_lazyown(method)lazyown.py:2689def run_lazyown(self)-- Run the internal module located atmodules/lazyown_parquet_tool.py. -
LazyOwnShell.run_update_db(method)lazyown.py:2715def run_update_db(self)-- Run the internal module located atmodules/update_db.sh. -
LazyOwnShell.run_lazynmap(method)lazyown.py:2745def run_lazynmap(self)-- Runs the internal modulemodules/lazynmap.shfor multiple Nmap scans. -
LazyOwnShell.run_lazywerkzeugdebug(method)lazyown.py:2810def run_lazywerkzeugdebug(self)-- Run the internal module located atmodules/legacy/lazywerkzeug.pyin debug mode. -
LazyOwnShell.run_lazygath(method)lazyown.py:2867def run_lazygath(self)-- Run the internal module located atmodules/lazygat.sh. to gathering the sistem :) -
LazyOwnShell.run_lazynmapdiscovery(method)lazyown.py:2899def run_lazynmapdiscovery(self)-- Runs the internal modulemodules/lazynmap.shwith discovery mode. -
LazyOwnShell.run_lazysniff(method)lazyown.py:2914def run_lazysniff(self)-- Run the sniffer internal module located atmodules/legacy/lazysniff.pywith the specified parameters. -
LazyOwnShell.run_lazyftpsniff(method)lazyown.py:2963def run_lazyftpsniff(self)-- Run the sniffer ftp internal module located atmodules/legacy/lazyftpsniff.pywith the specified parameters. -
LazyOwnShell.run_lazynetbios(method)lazyown.py:3009def run_lazynetbios(self)-- Run the internal module to search netbios vuln victims, located atmodules/legacy/lazynetbios.pywith the... -
LazyOwnShell.run_lazyhoneypot(method)lazyown.py:3060def run_lazyhoneypot(self)-- Run the internal module located atmodules/legacy/lazyhoneypot.pywith the specified parameters. -
LazyOwnShell.run_lazysearch_bot(method)lazyown.py:3126def run_lazysearch_bot(self)-- Run the internal module GROQ AI located atmodules/legacy/lazysearch_bot.pywith the specified parameters. -
LazyOwnShell.run_lazymetaextract0r(method)lazyown.py:3177def run_lazymetaextract0r(self)-- Run the Metadata extractor internal module located atmodules/lazyown_metaextract0r.pywith the specified parameters. -
LazyOwnShell.run_lazyownratcli(method)lazyown.py:3218def run_lazyownratcli(self)-- Run the internal module located atmodules/lazyownclient.pywith the specified parameters. -
LazyOwnShell.run_lazyownrat(method)lazyown.py:3278def run_lazyownrat(self)-- Run the internal module located atmodules/lazyownserver.pywith the specified parameters. -
LazyOwnShell.run_lazybotnet(method)lazyown.py:3339def run_lazybotnet(self)-- Run the internal module located atmodules/legacy/lazybotnet.pywith the specified parameters. -
LazyOwnShell.run_lazylfi2rce(method)lazyown.py:3395def run_lazylfi2rce(self)-- Run the internal module located atmodules/legacy/lazylfi2rce.pywith the specified parameters. -
LazyOwnShell.run_lazylogpoisoning(method)lazyown.py:3476def run_lazylogpoisoning(self)-- Run the internal module located atmodules/legacy/lazylogpoisoning.pywith the specified parameters. -
LazyOwnShell.run_lazybotcli(method)lazyown.py:3523def run_lazybotcli(self)-- Run the internal module located atmodules/legacy/lazybotcli.pywith the specified parameters. -
LazyOwnShell.run_lazyssh77enum(method)lazyown.py:3579def run_lazyssh77enum(self)-- Run the internal module located atmodules/lazybrutesshuserenum.pywith the specified parameters. -
LazyOwnShell.run_lazyburpfuzzer(method)lazyown.py:3629def run_lazyburpfuzzer(self)-- Run the internal module located atmodules/lazyown_burpfuzzer.pywith the specified parameters. -
LazyOwnShell.run_lazyreverse_shell(method)lazyown.py:3755def run_lazyreverse_shell(self)-- Run the internal module located atmodules/lazyreverse_shell.shwith the specified parameters. -
LazyOwnShell.run_lazyarpspoofing(method)lazyown.py:3805def run_lazyarpspoofing(self)-- Run the internal module located atmodules/legacy/lazyarpspoofing.pywith the specified parameters. -
LazyOwnShell.run_lazyattack(method)lazyown.py:3859def run_lazyattack(self)-- Run the internal module located atmodules/lazyatack.shwith the specified parameters. -
LazyOwnShell.run_lazymsfvenom(method)lazyown.py:3914def run_lazymsfvenom(self)-- Executes themsfvenomtool to generate a variety of payloads based on user input. -
LazyOwnShell.run_lazyaslrcheck(method)lazyown.py:4091def run_lazyaslrcheck(self)-- Creates a path hijacking attack by performing the following steps: -
LazyOwnShell.run_lazypathhijacking(method)lazyown.py:4138def run_lazypathhijacking(self)-- Creates a path hijacking attack by performing the following steps: -
LazyOwnShell.run_script(method)lazyown.py:4171def run_script(self, script_name)-- Run a script with the given arguments -
LazyOwnShell.run_command(method)lazyown.py:4203def run_command(self, command)-- Run a command and print output in real-time -
LazyOwnShell.get_output(method)lazyown.py:4283def get_output(self)-- Devuelve la salida acumulada -
LazyOwnShell.upload_file_to_c2(method)lazyown.py:4288def upload_file_to_c2(self, file_path, clientid)-- Sube un archivo al C2. -
LazyOwnShell.complete_upload_c2(method)lazyown.py:4314def complete_upload_c2(self, text, line, begidx, endidx)-- Autocomplete implant names from implant_config_*.json files in sessions/ directory -
LazyOwnShell.download_file_from_c2(method)lazyown.py:4342def download_file_from_c2(self, file_name, clientid)-- Descarga un archivo desde el C2. -
LazyOwnShell.issue_command_to_c2(method)lazyown.py:4411def issue_command_to_c2(self, command, client_id)-- Ejecuta un comando en el cliente usando el C2. -
LazyOwnShell.complete_issue_command_to_c2(method)lazyown.py:4445def complete_issue_command_to_c2(self, text, line, begidx, endidx)-- Autocomplete: 1st arg = implant name, 2nd arg = beacon command (with : if needed) -
LazyOwnShell.view_code(method)lazyown.py:4511def view_code(self, stdscr)-- Display C and ASM code side by side in a curses-based interface. -
LazyOwnShell.get_available_actions(method)lazyown.py:4616def get_available_actions(self)-- Returns a list of available actions using cmd2 introspection. -
LazyOwnShell.process_scan_csv(method)lazyown.py:4699def process_scan_csv(self, csv_file, ip, port, all_data, processed_ips)-- Processes a single scan CSV file. -
LazyOwnShell.process_vuln_csv(method)lazyown.py:4722def process_vuln_csv(self, csv_file, ip, all_data, processed_ips)-- Processes a single vulnerability CSV file. -
LazyOwnShell.do_event_log(method)lazyown.py:4854def do_event_log(self, line)-- Show recent EventBus events. -
LazyOwnShell.do_state_snapshot(method)lazyown.py:4884def do_state_snapshot(self, line)-- Show unified StateManager snapshot (DB + JSON caches). -
LazyOwnShell.do_route(method)lazyown.py:4923def do_route(self, line)-- Route a natural-language prompt to a LazyOwn tool. -
main(function)lazyown.py:4946def main()
-
auth(function)modules/49803.py:43def auth() -
injection(function)modules/49803.py:69def injection() -
connection(function)modules/49803.py:106def connection()
-
poc(function)modules/CVE-2023-28432.py:12def poc(url)
-
AutocompleteEntry.__init__(method)modules/LazyOwnExplorer.py:31def __init__(self, get_suggestions_func) -
AutocompleteEntry.changed(method)modules/LazyOwnExplorer.py:45def changed(self, name, index, mode) -
AutocompleteEntry.selection(method)modules/LazyOwnExplorer.py:68def selection(self, event) -
AutocompleteEntry.move_up(method)modules/LazyOwnExplorer.py:75def move_up(self, event) -
AutocompleteEntry.move_down(method)modules/LazyOwnExplorer.py:87def move_down(self, event) -
AutocompleteEntry.comparison(method)modules/LazyOwnExplorer.py:99def comparison(self, pattern) -
LazyOwnGUI.__init__(method)modules/LazyOwnExplorer.py:105def __init__(self) -
LazyOwnGUI.create_widgets(method)modules/LazyOwnExplorer.py:119def create_widgets(self) -
get_suggestions(function)modules/LazyOwnExplorer.py:124def get_suggestions(term) -
LazyOwnGUI.load_parquet_files(method)modules/LazyOwnExplorer.py:181def load_parquet_files(self) -
LazyOwnGUI.get_unique_values(method)modules/LazyOwnExplorer.py:241def get_unique_values(self) -
LazyOwnGUI.search(method)modules/LazyOwnExplorer.py:248def search(self) -
LazyOwnGUI.search_in_parquet(method)modules/LazyOwnExplorer.py:267def search_in_parquet(self, term) -
LazyOwnGUI.on_row_double_click(method)modules/LazyOwnExplorer.py:278def on_row_double_click(self, event) -
LazyOwnGUI.show_row_details(method)modules/LazyOwnExplorer.py:283def show_row_details(self, row) -
LazyOwnGUI.add_new_attack_vector(method)modules/LazyOwnExplorer.py:305def add_new_attack_vector(self) -
save_new_vector(function)modules/LazyOwnExplorer.py:325def save_new_vector() -
LazyOwnGUI.scan_system_for_binaries(method)modules/LazyOwnExplorer.py:357def scan_system_for_binaries(self) -
is_binary(function)modules/LazyOwnExplorer.py:358def is_binary(file_path) -
LazyOwnGUI.show_scan_results(method)modules/LazyOwnExplorer.py:374def show_scan_results(self, binaries) -
LazyOwnGUI.export_to_csv(method)modules/LazyOwnExplorer.py:385def export_to_csv(self)
Imported by: cli/commands/exploit_migrated.py
-
CertificateTemplate.esc_vulnerabilities(method)modules/adcs_attacks.py:35def esc_vulnerabilities(self)-- Determine which ESC attack paths apply to this template. -
ADCSCertipyWrapper.__init__(method)modules/adcs_attacks.py:80def __init__(self, certipy_path, timeout) -
ADCSCertipyWrapper.find_certificate_authorities(method)modules/adcs_attacks.py:113def find_certificate_authorities(self, username, password, domain, dc_ip, hashes)-- Enumerate certificate authorities in an Active Directory domain. -
ADCSCertipyWrapper.enumerate_templates(method)modules/adcs_attacks.py:189def enumerate_templates(self, username, password, domain, dc_ip, hashes)-- Enumerate vulnerable certificate templates. -
ADCSCertipyWrapper.request_certificate_esc1(method)modules/adcs_attacks.py:268def request_certificate_esc1(self, username, password, domain, dc_ip, ca_name, template_name, target_user, output_file)-- ESC1: Request a certificate with a user-supplied subject alternative name. -
ADCSCertipyWrapper.request_certificate_esc8(method)modules/adcs_attacks.py:322def request_certificate_esc8(self, username, password, domain, dc_ip, ca_server, template_name, output_file)-- ESC8: HTTP-based certificate enrollment (NTLM relay to AD CS). -
ADCSCertipyWrapper.authenticate_with_certificate(method)modules/adcs_attacks.py:367def authenticate_with_certificate(self, cert_file, domain, dc_ip, username)-- Authenticate to the domain using a certificate and retrieve NT hash. -
ADCSCertipyWrapper.assess_vulnerability(method)modules/adcs_attacks.py:411def assess_vulnerability(self, username, password, domain, dc_ip, hashes)-- Run a full AD CS vulnerability assessment.
Depends on: core/logging.py
Imported by: cli/commands/lazyops.py, cli/commands/marketplace.py, tests/test_addon_updater.py
-
AddonUpdateResult.to_dict(method)modules/addon_updater.py:46def to_dict(self)-- Return a JSON-serializable view. -
load_addons(function)modules/addon_updater.py:73def load_addons(lazyaddons_dir)-- Load every lazyaddon with its update metadata. -
update_addon(function)modules/addon_updater.py:166def update_addon(name, lazyaddons_dir, root, dry_run)-- Update one addon throughgit pull --ff-only. -
update_all(function)modules/addon_updater.py:260def update_all(names, lazyaddons_dir, root, dry_run, audit_path)-- Update every opted-in addon or an explicit name list.
Depends on: core/logging.py, modules/ai_model.py, modules/llm_factory.py, modules/logging_config.py
Imported by: modules/lazy_agent.py, modules/vulnbot.py, tests/test_lazy_streaming.py
-
configure_logging(function)modules/agent_runner.py:60def configure_logging(debug) -
AgentTool.__init__(method)modules/agent_runner.py:69def __init__(self, name, description, func, parameters, required) -
AgentTool.to_api_format(method)modules/agent_runner.py:77def to_api_format(self) -
AgentTool.execute(method)modules/agent_runner.py:83def execute(self)-- Execute with argument validation and clear formatting. -
ASTToolExtractor.extract_commands_from_file(method)modules/agent_runner.py:138def extract_commands_from_file(file_path, prefix) -
AgentRunner.__init__(method)modules/agent_runner.py:193def __init__(self, model, system_prompt, max_iterations) -
AgentRunner.register_tool(method)modules/agent_runner.py:218def register_tool(self, tool) -
AgentRunner.register_tool_from_instance(method)modules/agent_runner.py:221def register_tool_from_instance(self, func)-- Wrap a plain function as an AgentTool using its signature and docstring. -
AgentRunner.register_tools_from_metadata(method)modules/agent_runner.py:239def register_tools_from_metadata(self, commands, executor) -
make_executor(function)modules/agent_runner.py:243def make_executor(cmd_name) -
wrapper(function)modules/agent_runner.py:244def wrapper(command) -
AgentRunner.get_tools_for_api(method)modules/agent_runner.py:266def get_tools_for_api(self) -
AgentRunner.run(method)modules/agent_runner.py:271def run(self, user_input) -
LazyOwnShellWrapper.__init__(method)modules/agent_runner.py:485def __init__(self, script_path) -
LazyOwnShellWrapper.execute_command(method)modules/agent_runner.py:515def execute_command(self, command)-- Execute a command with a timeout using threading. -
target(function)modules/agent_runner.py:522def target() -
LazyOwnShellWrapper.get_commands_summary(method)modules/agent_runner.py:552def get_commands_summary(self) -
VulnBotCLI.__init__(method)modules/agent_runner.py:560def __init__(self, provider, mode, debug, script_path) -
VulnBotCLI.process_request(method)modules/agent_runner.py:615def process_request(self, user_input) -
interactive_mode(function)modules/agent_runner.py:620def interactive_mode(bot) -
parse_args(function)modules/agent_runner.py:644def parse_args() -
main(function)modules/agent_runner.py:654def main()
Depends on: modules/autonomous_exploit_engine.py
Imported by: cli/commands/pwn.py, skills/lazyown_mcp.py
-
AIExploitChainer.__init__(method)modules/ai_exploit_chain.py:259def __init__(self) -
AIExploitChainer.reason(method)modules/ai_exploit_chain.py:285def reason(self, context)-- Analyze context failures and produce the next best exploit candidate. -
AIExploitChainer.evaluate_failure(method)modules/ai_exploit_chain.py:306def evaluate_failure(result)-- Classify the reason for an exploit failure by output pattern matching. -
AIExploitChainer.select_next_strategy(method)modules/ai_exploit_chain.py:318def select_next_strategy(self, context)-- Select the next strategy to attempt based on context and failures. -
AIExploitChainer.estimate_confidence(method)modules/ai_exploit_chain.py:349def estimate_confidence(self, strategy, profile)-- Estimate confidence (0.0-1.0) for a strategy against a target profile. -
AIExploitChainer.build_chain_plan(method)modules/ai_exploit_chain.py:391def build_chain_plan(self, context)-- Build a full multi-step exploitation plan following the kill chain. -
AIExploitChainer.adapt_chain(method)modules/ai_exploit_chain.py:428def adapt_chain(self, context, new_info)-- Adapt the chain plan based on newly discovered information.
Depends on: modules/llm_factory.py, modules/toposwarm_bridge.py
Imported by: modules/recommender.py, modules/timeline_narrator.py
-
call(function)modules/ai_fallback.py:254def call(prompt, system, api_key, max_tokens, temperature)-- Call Groq → fallback to Ollama → fallback to TopoSwarm local brain → error.
Depends on: core/logging.py, modules/llm_factory.py
Imported by: contrib/legacy/lazyllmchat.py, modules/agent_runner.py, modules/llm_client.py, modules/llm_factory.py, modules/vulnbot.py, modules/yaml_generator.py
-
AIModel.generate(method)modules/ai_model.py:56def generate(self, prompt)-- Return a non-streaming completion forprompt. -
AIModel.stream_generate(method)modules/ai_model.py:60def stream_generate(self, prompt)-- Yield streaming completion chunks forprompt. -
AIModel.complete(method)modules/ai_model.py:63def complete(self, system, user, max_tokens, temperature)-- Return a completion conforming tocore.protocols.LLMBackend. -
_LazyImporter.groq(method)modules/ai_model.py:105def groq(cls) -
_LazyImporter.openai(method)modules/ai_model.py:113def openai(cls) -
_LazyImporter.anthropic(method)modules/ai_model.py:121def anthropic(cls) -
GroqModel.__init__(method)modules/ai_model.py:137def __init__(self, api_key, model) -
GroqModel.generate(method)modules/ai_model.py:155def generate(self, prompt)-- Return a single completion forprompt. -
GroqModel.stream_generate(method)modules/ai_model.py:183def stream_generate(self, prompt)-- Yield streamed completion chunks forprompt. -
GroqModel.complete(method)modules/ai_model.py:216def complete(self, system, user, max_tokens, temperature)-- Return a chat completion using native role separation. -
OllamaModel.__init__(method)modules/ai_model.py:252def __init__(self, model, host) -
OllamaModel.generate(method)modules/ai_model.py:260def generate(self, prompt)-- Return a single completion forprompt. -
OllamaModel.stream_generate(method)modules/ai_model.py:281def stream_generate(self, prompt)-- Yield streamed completion chunks forprompt. -
LlamaCppModel.__init__(method)modules/ai_model.py:324def __init__(self, model, host) -
LlamaCppModel.generate(method)modules/ai_model.py:362def generate(self, prompt)-- Return a single completion forprompt. -
LlamaCppModel.stream_generate(method)modules/ai_model.py:382def stream_generate(self, prompt)-- Yield streamed completion deltas forprompt. -
LlamaCppModel.complete(method)modules/ai_model.py:413def complete(self, system, user, max_tokens, temperature)-- Role-aware completion through the chat endpoint. -
OpenAIModel.__init__(method)modules/ai_model.py:440def __init__(self, api_key, model) -
OpenAIModel.generate(method)modules/ai_model.py:444def generate(self, prompt) -
OpenAIModel.stream_generate(method)modules/ai_model.py:455def stream_generate(self, prompt) -
OpenAIModel.complete(method)modules/ai_model.py:470def complete(self, system, user, max_tokens, temperature) -
AnthropicModel.__init__(method)modules/ai_model.py:499def __init__(self, api_key, model) -
AnthropicModel.generate(method)modules/ai_model.py:503def generate(self, prompt) -
AnthropicModel.stream_generate(method)modules/ai_model.py:515def stream_generate(self, prompt) -
AnthropicModel.complete(method)modules/ai_model.py:529def complete(self, system, user, max_tokens, temperature) -
DeepSeekModel.__init__(method)modules/ai_model.py:556def __init__(self, api_key, model) -
DeepSeekModel.generate(method)modules/ai_model.py:563def generate(self, prompt) -
DeepSeekModel.stream_generate(method)modules/ai_model.py:574def stream_generate(self, prompt) -
DeepSeekModel.complete(method)modules/ai_model.py:589def complete(self, system, user, max_tokens, temperature)
-
LoadNtFunctions(function)modules/amsi.c:38void LoadNtFunctions()-- Función para cargar las funciones dinámicamente -
AMS1patch_OpenSession_jne(function)modules/amsi.c:66void AMS1patch_OpenSession_jne(HANDLE hproc)-- Técnica de Saad, también conocido como @D1rkMtr (https://twitter.com/D1rkMtr/)... -
AMS1patch_OpenSession_ret(function)modules/amsi.c:116void AMS1patch_OpenSession_ret(HANDLE hproc) -
AMS1patch_ScanBuffer_ret(function)modules/amsi.c:166void AMS1patch_ScanBuffer_ret(HANDLE hproc) -
AMS1patch_RastaMouse(function)modules/amsi.c:215void AMS1patch_RastaMouse(HANDLE hproc) -
AMS1patch_E_ACCESSDENIED(function)modules/amsi.c:274void AMS1patch_E_ACCESSDENIED(HANDLE hproc) -
AMS1patch_E_HANDLE(function)modules/amsi.c:326void AMS1patch_E_HANDLE(HANDLE hproc) -
AMS1patch_E_OUTOFMEMORY(function)modules/amsi.c:378void AMS1patch_E_OUTOFMEMORY(HANDLE hproc) -
main(function)modules/amsi.c:428int main(int argc, char** argv)
-
start(function)modules/amt_auth_bypass.py:4def start() -
BlankAuthResponse.request(method)modules/amt_auth_bypass.py:11def request(self, flow)
Depends on: core/console.py, modules/atomic_resolver.py
Imported by: cli/commands/command_and_control_migrated.py, cli/orchestrator_assets_tabs.py, cli/recommendation_signals.py, modules/emulation_runner.py, modules/operation.py, tests/test_atomic_loop.py, tests/test_emulation_runner.py
-
AptPlaybook.from_dict(method)modules/apt_playbooks.py:74def from_dict(cls, data) -
AptPlaybook.to_summary(method)modules/apt_playbooks.py:99def to_summary(self) -
AptPlaybookEngine.__init__(method)modules/apt_playbooks.py:114def __init__(self, playbook_dir) -
AptPlaybookEngine.list_playbooks(method)modules/apt_playbooks.py:131def list_playbooks(self) -
AptPlaybookEngine.get(method)modules/apt_playbooks.py:134def get(self, name) -
AptPlaybookEngine.validate(method)modules/apt_playbooks.py:138def validate(self, pb, atomic_path)-- Validate that all referenced atomic tests exist locally. -
AptPlaybookEngine.generate_attack_plan(method)modules/apt_playbooks.py:213def generate_attack_plan(self, pb, output_path)-- Generate a legacyattack_plan.yamlfrom an enriched APT playbook. -
AptPlaybookEngine.report_json(method)modules/apt_playbooks.py:230def report_json(self, pb, output_path)-- Write a structured JSON report of the playbook for the operator.
Imported by: skills/autonomous_daemon.py, tests/test_autonomous_ops.py
-
asset_key(function)modules/asset_graph.py:28def asset_key(asset_type, value, parent_key)-- Compute the stable deduplication key for an asset. -
AssetGraph.__init__(method)modules/asset_graph.py:73def __init__(self, config)-- Initialize empty graphs and anchor indexes. -
AssetGraph.upsert_asset(method)modules/asset_graph.py:82def upsert_asset(self, asset_type, value, parent_key, company)-- Insert or refresh an asset and return the canonical node. -
AssetGraph.add_node(method)modules/asset_graph.py:108def add_node(self, task_id, kind, text, parent_id, payload)-- Append an exploration node to a task chain. -
AssetGraph.add_fact(method)modules/asset_graph.py:123def add_fact(self, task_id, text, parent_id, payload)-- Append a fact node to a task chain. -
AssetGraph.add_finding(method)modules/asset_graph.py:128def add_finding(self, task_id, text, parent_id, payload)-- Append a finding node to a task chain. -
AssetGraph.anchor(method)modules/asset_graph.py:133def anchor(self, exploration_id, key)-- Link an exploration node to an asset key. -
AssetGraph.assets_for_node(method)modules/asset_graph.py:145def assets_for_node(self, exploration_id)-- List assets anchored to one exploration node. -
AssetGraph.nodes_for_asset(method)modules/asset_graph.py:151def nodes_for_asset(self, key)-- List exploration nodes anchored to one asset. -
AssetGraph.coverage(method)modules/asset_graph.py:157def coverage(self, task_id)-- Report tested versus total assets for a task. -
AssetGraph.mirror_to_campaign(method)modules/asset_graph.py:170def mirror_to_campaign(self, db, world_model, workspace_id)-- Mirror graph assets into the campaign database and world model. -
coverage_for(function)modules/asset_graph.py:214def coverage_for(db, workspace_id, scope)-- Report tested scope assets versus total scope from the campaign db.
Depends on: core/console.py, core/hardening.py, core/logging.py, modules/atomic_resolver.py, utils.py
Imported by: cli/commands/command_and_control_migrated.py
-
parse_atomic_line(function)modules/atomic_cli.py:37def parse_atomic_line(line)-- Parse free-form CLI args into a structured request. -
run_atomic_tests(function)modules/atomic_cli.py:185def run_atomic_tests(params, line, shell)-- Execute the atomic_tests workflow (secure + agent-safe). -
run_atomic_gen(function)modules/atomic_cli.py:325def run_atomic_gen(params, line, shell)-- Generate test + cleanup scripts for one technique or GUID.
Depends on: core/logging.py
Imported by: skills/lazyown_groq_agents.py, skills/lazyown_mcp.py, skills/lazyown_parquet_db.py, tests/test_core_modules.py
-
enrich(function)modules/atomic_enricher.py:173def enrich(force)-- Build enriched parquet. -
load_enriched(function)modules/atomic_enricher.py:215def load_enriched()-- Load enriched parquet, building it first if it doesn't exist. -
query_atomic(function)modules/atomic_enricher.py:232def query_atomic(keyword, mitre_id, platform, scope, has_prereqs, complexity, limit, include_command)-- Structured query over the enriched Atomic Red Team technique catalogue.
Depends on: core/logging.py, core/safe_exec.py, modules/c2_surfaces.py, modules/emulation_runner.py, modules/outcome_bus.py, modules/world_model.py
Imported by: cli/commands/command_and_control_migrated.py, modules/autonomous_exploit_engine.py, modules/c2_agent.py, modules/emulation_runner.py, modules/metasploitable_gym.py, modules/redteam_gym.py, skills/lazyown_mcp.py, tests/test_atomic_exploit_bridge.py, tests/test_c2_surfaces.py, tests/test_emulation_runner.py, tests/test_gym_cli.py, tests/test_metasploitable_gym.py, tests/test_outcome_bus.py
-
ExploitSpec.severity_rank(method)modules/atomic_exploit_bridge.py:103def severity_rank(self)-- Return numeric severity for ranking (mirrors the engine). -
ExploitSpec.to_dict(method)modules/atomic_exploit_bridge.py:107def to_dict(self)-- Return a JSON-serializable view. -
validate_spec(function)modules/atomic_exploit_bridge.py:136def validate_spec(data, filename)-- Validate raw YAML data against the exploit schema. -
load_all(function)modules/atomic_exploit_bridge.py:212def load_all(exploits_dir)-- Load every enabled, valid exploit YAML (skips the rest loudly). -
match_score(function)modules/atomic_exploit_bridge.py:267def match_score(spec, service, product, version)-- Score a spec against one discovered service (0 when no match). -
parse_scan_xml(function)modules/atomic_exploit_bridge.py:286def parse_scan_xml(xml_path)-- Parse open TCP ports from a lazynmap XML file (stdlib only). -
auto(function)modules/atomic_exploit_bridge.py:323def auto(target, lhost, max_attempts, stop_on_shell, task_id, exploits_dir, scan_dir)-- Pwntomate-style sweep: triggers from scan XML, ranked, stop on SHELL. -
rank(function)modules/atomic_exploit_bridge.py:421def rank(services, exploits_dir)-- Rank every spec against discovered services (severity, then confidence). -
build_commands(function)modules/atomic_exploit_bridge.py:452def build_commands(spec, rhost, lhost, lport, port)-- Substitute{tokens}into a spec's commands (no shell involved). -
detect_shell(function)modules/atomic_exploit_bridge.py:470def detect_shell(spec, output)-- Return True when output carries any shell indicator. -
run_spec(function)modules/atomic_exploit_bridge.py:512def run_spec(spec, rhost, lhost, lport, port)-- Execute a spec's commands in order; stop at the first SHELL hit. -
open_shell(function)modules/atomic_exploit_bridge.py:592def open_shell(target, services, lhost, exploits_dir, attempts)-- Return the first exec-capable spec with a live multiplexed session. -
_ShellSession.__init__(method)modules/atomic_exploit_bridge.py:635def __init__(self, target, port, sock, owner) -
_ShellSession.alive(method)modules/atomic_exploit_bridge.py:645def alive(self, timeout)-- Return True when the session answers a no-op probe. -
_ShellSession.run(method)modules/atomic_exploit_bridge.py:652def run(self, command, timeout)-- Run one command, return stdout+stderr (never raises). -
_ShellSession.close(method)modules/atomic_exploit_bridge.py:694def close(self)-- End the remote shell cleanly so its listener dies too. -
record_shell_outcome(function)modules/atomic_exploit_bridge.py:745def record_shell_outcome(target, via)-- Mark one target as compromised after an operator obtains a shell. -
live_sessions(function)modules/atomic_exploit_bridge.py:774def live_sessions()-- Return in-process shell sessions without probing the target. -
close_sessions(function)modules/atomic_exploit_bridge.py:793def close_sessions(target)-- Close cached sessions (all, or one target). -
cleanup_target(function)modules/atomic_exploit_bridge.py:831def cleanup_target(target, lhost, hard, gym_container, exploits_dir, staging_dir)-- Clean up after an engagement (soft default, hard on request). -
ensure_session(function)modules/atomic_exploit_bridge.py:907def ensure_session(spec, target, lhost)-- Return a live multiplexed session, establishing it when needed. -
remote_exec(function)modules/atomic_exploit_bridge.py:1003def remote_exec(spec, target, command, lhost)-- Run one command through the persistent multiplexed session. -
interactive_shell(function)modules/atomic_exploit_bridge.py:1021def interactive_shell(target, lhost, interactive, exploits_dir)-- Open a multiplexed session; drop into a REPL only when allowed. -
emulate(function)modules/atomic_exploit_bridge.py:1076def emulate(target, techniques, playbook, platform, lhost, per_technique, exploits_dir, scan_dir, include_manual...-- Execute emulation tests on the target through one working shell. -
record_outcome(function)modules/atomic_exploit_bridge.py:1110def record_outcome(spec, target, shell, output, task_id, intent_id, surfaces_dir, port, service, product, version...-- Feed one exploit outcome to every shared surface through the bus. -
main(function)modules/atomic_exploit_bridge.py:1198def main()-- CLI: validate | list | rank | run (used by the addon and operators).
Depends on: core/logging.py, core/safe_exec.py
Imported by: cli/commands/command_and_control_migrated.py, modules/apt_playbooks.py, modules/atomic_cli.py, modules/emulation_runner.py, modules/ttpscope_bridge.py, skills/lazyown_mcp.py, tests/test_atomic_loop.py, tests/test_atomic_resolver.py
-
normalize_mitre_id(function)modules/atomic_resolver.py:66def normalize_mitre_id(raw)-- Normalize and validate a MITRE technique ID. -
is_mitre_id(function)modules/atomic_resolver.py:84def is_mitre_id(raw)-- Return True when raw looks like a MITRE technique ID. -
normalize_guid(function)modules/atomic_resolver.py:93def normalize_guid(raw)-- Normalize and validate an Atomic test GUID. -
is_guid(function)modules/atomic_resolver.py:109def is_guid(raw)-- Return True when raw looks like an Atomic test GUID. -
normalize_platform(function)modules/atomic_resolver.py:118def normalize_platform(raw)-- Validate a platform against the allowlist. -
AtomicTest.mitre_url(method)modules/atomic_resolver.py:147def mitre_url(self)-- Return the ATT&CK URL for this test. -
AtomicTest.to_dict(method)modules/atomic_resolver.py:151def to_dict(self)-- Return a JSON-serializable view. -
AtomicResolver.__init__(method)modules/atomic_resolver.py:226def __init__(self, atomics_dir) -
AtomicResolver.available(method)modules/atomic_resolver.py:233def available(self)-- Return True when the atomics directory exists. -
AtomicResolver.ensure_repo(method)modules/atomic_resolver.py:237def ensure_repo(self)-- Clone or pull the Atomic Red Team repo without shell. -
AtomicResolver.build_index(method)modules/atomic_resolver.py:265def build_index(self, force)-- Build the MITRE -> [GUID] index (cached unless force). -
AtomicResolver.guids_for_technique(method)modules/atomic_resolver.py:305def guids_for_technique(self, mitre_id)-- Return GUIDs registered for a technique ID (empty when unknown). -
AtomicResolver.get_by_guid(method)modules/atomic_resolver.py:311def get_by_guid(self, guid)-- Return the test for a GUID, or None when unknown. -
AtomicResolver.list_for_technique(method)modules/atomic_resolver.py:317def list_for_technique(self, mitre_id, platform)-- List tests for a technique, optionally filtered by platform. -
AtomicResolver.resolve(method)modules/atomic_resolver.py:334def resolve(self, query, platform)-- Resolve a MITRE ID, GUID, or prefix to matching tests. -
AtomicResolver.chain(method)modules/atomic_resolver.py:380def chain(self, techniques, platform, first_match_only)-- Translate an ordered APT technique chain to concrete tests. -
AtomicResolver.stats(method)modules/atomic_resolver.py:415def stats(self)-- Return index counters. -
get_resolver(function)modules/atomic_resolver.py:424def get_resolver(atomics_dir)-- Return the module-level resolver singleton. -
resolve_technique(function)modules/atomic_resolver.py:432def resolve_technique(mitre_id, platform, atomics_dir)-- Resolve a technique to JSON-serializable test dicts. -
resolve_chain(function)modules/atomic_resolver.py:442def resolve_chain(techniques, platform, atomics_dir)-- Translate a technique chain to JSON-serializable test dicts. -
build_variables(function)modules/atomic_resolver.py:452def build_variables(test, rhost, domain, rport, username, password)-- Build #{variable} substitution map for a test. -
render_command(function)modules/atomic_resolver.py:488def render_command(template, variables)-- Substitute #{variables} without shell interpretation. -
technique_env_override(function)modules/atomic_resolver.py:496def technique_env_override()-- Return atomics dir override from env (empty when unset).
Depends on: cli/surface_graph.py, modules/estorides_fusion.py, modules/finding_lifecycle.py, modules/intel_fusion.py
Imported by: cli/commands/surface.py, lazyc2.py, modules/sinks/attack_graph.py, tests/test_attack_surface3d.py
-
build_attack_graph(function)modules/attack_surface3d.py:147def build_attack_graph(sessions_dir, payload_path, est_cap)-- Fuse surface scans with campaign intel into one C2-centred graph. -
add_edge(function)modules/attack_surface3d.py:155def add_edge(src, dst, kind, evidence, weight)-- Append one provenance-tagged relation between known nodes. -
tag(function)modules/attack_surface3d.py:166def tag(nid, source)-- Record a corroborating artefact on a node, deduped. -
refresh_summary(function)modules/attack_surface3d.py:460def refresh_summary(graph, sessions_dir)-- Score exposure into intel_summary.json for the autonomous loop.
Imported by: skills/lazyown_mcp.py, tests/test_auto_loop_memory.py
-
FailureMemory.__init__(method)modules/auto_loop_memory.py:56def __init__(self, filepath) -
FailureMemory.path(method)modules/auto_loop_memory.py:63def path(self)-- Backing file path for administrative inspection. -
FailureMemory.make_key(method)modules/auto_loop_memory.py:68def make_key(target, category, command)-- Build the storage key; raises on empty target or command. -
FailureMemory.base_of(method)modules/auto_loop_memory.py:75def base_of(command)-- Extract the base command from a full command string. -
FailureMemory.record_failure(method)modules/auto_loop_memory.py:79def record_failure(self, target, category, command)-- Increment and persist the failure count; returns the new count. -
FailureMemory.failure_count(method)modules/auto_loop_memory.py:89def failure_count(self, target, category, command)-- Current persisted failure count for the combination. -
FailureMemory.count_for_base(method)modules/auto_loop_memory.py:94def count_for_base(self, target, base_command)-- Failure hits aggregated across exact keys whose base matches. -
FailureMemory.veto_base(method)modules/auto_loop_memory.py:105def veto_base(self, target, base_command)-- Mark a base command as blocked by the stuck-loop detector. -
FailureMemory.is_blocked(method)modules/auto_loop_memory.py:115def is_blocked(self, target, base_command, bonus_key)-- True when the base is vetoed or its failure count hit the threshold. -
FailureMemory.failure_count_strict(method)modules/auto_loop_memory.py:123def failure_count_strict(self, target, key)-- Failure count for a raw storage key already containing separators. -
FailureMemory.blocked_bases(method)modules/auto_loop_memory.py:128def blocked_bases(self, target)-- All blocked base commands for the target (vetoed plus past threshold). -
FailureMemory.clear_target(method)modules/auto_loop_memory.py:140def clear_target(self, target)-- Forget every record for a target (manual reset path). -
FailureMemory.sanctioned_bases(method)modules/auto_loop_memory.py:146def sanctioned_bases(self, target)-- Administrative view: counts above threshold with their last timestamp. -
FailureMemory.save(method)modules/auto_loop_memory.py:166def save(self) -
DecisionPersistence.__init__(method)modules/auto_loop_memory.py:215def __init__(self, filepath) -
DecisionPersistence.push(method)modules/auto_loop_memory.py:220def push(self, command, reason)-- Park one decision for later consumption (newest first). -
DecisionPersistence.pop(method)modules/auto_loop_memory.py:229def pop(self)-- Consume and persist-removal of the newest parked decision. -
DecisionPersistence.peek(method)modules/auto_loop_memory.py:237def peek(self)-- Newest parked decision without consuming it. -
DecisionPersistence.save(method)modules/auto_loop_memory.py:244def save(self)
Imported by: modules/autonomous_exploit_engine.py, skills/lazyown_mcp.py
-
AutoPivotEngine.__init__(method)modules/auto_pivot.py:73def __init__(self, config) -
AutoPivotEngine.set_config(method)modules/auto_pivot.py:82def set_config(self, config) -
AutoPivotEngine.add_pivot_node(method)modules/auto_pivot.py:90def add_pivot_node(self, ip, access_type, credential_id, hostname) -
AutoPivotEngine.remove_pivot_node(method)modules/auto_pivot.py:120def remove_pivot_node(self, ip) -
AutoPivotEngine.discover_subnets(method)modules/auto_pivot.py:128def discover_subnets(self, node_ip, command_output) -
AutoPivotEngine.build_chain(method)modules/auto_pivot.py:146def build_chain(self, target_ip, via_nodes, subnet) -
AutoPivotEngine.generate_proxychains_config(method)modules/auto_pivot.py:177def generate_proxychains_config(self, chain_id) -
AutoPivotEngine.generate_sshuttle_command(method)modules/auto_pivot.py:203def generate_sshuttle_command(self, node_ip, subnet) -
AutoPivotEngine.generate_chisel_command(method)modules/auto_pivot.py:206def generate_chisel_command(self, target_ip, local_port) -
AutoPivotEngine.generate_ssh_socks_command(method)modules/auto_pivot.py:212def generate_ssh_socks_command(self, node_ip, local_port) -
AutoPivotEngine.health_check(method)modules/auto_pivot.py:216def health_check(self, node_ip) -
AutoPivotEngine.health_check_all(method)modules/auto_pivot.py:237def health_check_all(self) -
AutoPivotEngine.get_active_routes(method)modules/auto_pivot.py:243def get_active_routes(self) -
AutoPivotEngine.get_reachable_hosts(method)modules/auto_pivot.py:259def get_reachable_hosts(self, node_ip) -
AutoPivotEngine.set_reachable_hosts(method)modules/auto_pivot.py:267def set_reachable_hosts(self, node_ip, hosts) -
AutoPivotEngine.set_route_priority(method)modules/auto_pivot.py:272def set_route_priority(self, node_ip, priority) -
AutoPivotEngine.suggest_next_pivot(method)modules/auto_pivot.py:352def suggest_next_pivot(self)
Depends on: core/config.py, core/logging.py, core/process.py, modules/detection_oracle.py
Imported by: cli/commands/purple_team.py
-
PurpleResult.to_csv_row(method)modules/auto_purple.py:146def to_csv_row(self)-- Serialize to a flat dict suitable for CSV writer. -
PurpleResult.to_dict(method)modules/auto_purple.py:160def to_dict(self) -
IPurpleLoop.execute_and_measure(method)modules/auto_purple.py:185def execute_and_measure(self, command, args, category) -
IPurpleLoop.measure_only(method)modules/auto_purple.py:193def measure_only(self, command, args, category) -
IPurpleLoop.engagement_score(method)modules/auto_purple.py:201def engagement_score(self) -
IPurpleLoop.export_dataset(method)modules/auto_purple.py:204def export_dataset(self) -
IPurpleLoop.export_report(method)modules/auto_purple.py:207def export_report(self) -
IPurpleEvaluator.last_results(method)modules/auto_purple.py:214def last_results(self, n) -
IPurpleEvaluator.category_accuracy(method)modules/auto_purple.py:217def category_accuracy(self, category) -
PurpleTeamLoop.__init__(method)modules/auto_purple.py:414def __init__(self, bt_path, detection_delay, methods, auto_feedback, session_id) -
PurpleTeamLoop.execute_and_measure(method)modules/auto_purple.py:443def execute_and_measure(self, command, args, category)-- Execute a red action and measure if LazyOwnBT detects it. -
PurpleTeamLoop.measure_only(method)modules/auto_purple.py:493def measure_only(self, command, args, category)-- Measure without executing — test if BT would detect the command. -
PurpleTeamLoop.engagement_score(method)modules/auto_purple.py:529def engagement_score(self)-- Aggregate detection score from all results in this session. -
PurpleTeamLoop.export_dataset(method)modules/auto_purple.py:567def export_dataset(self)-- Return the path to the CSV dataset file. -
PurpleTeamLoop.export_report(method)modules/auto_purple.py:571def export_report(self)-- Generate a full engagement report. -
PurpleTeamLoop.last_results(method)modules/auto_purple.py:596def last_results(self, n) -
PurpleTeamLoop.category_accuracy(method)modules/auto_purple.py:599def category_accuracy(self, category) -
get_loop(function)modules/auto_purple.py:772def get_loop(bt_path, detection_delay, methods, auto_feedback)-- Return (or create) the module-level PurpleTeamLoop singleton. -
load_config_from_payload(function)modules/auto_purple.py:790def load_config_from_payload(params)-- Extract purple_team config from payload.json params dict.
Next: API_p10.md
Generated offline by readmenator from commit 3572cf8f9c41. Edits here are overwritten by readmenator . --rebuild.
Start
Wiki
- community_0_modules_autonomous_daemon
- community_10_skills_claude_md_orchestrator
- community_11_core
- community_12_modules_finding_lifecycle
- community_13_modules_kerberos_core
- community_14_modules_saas_attacks
- community_15_modules_polymorphic_engine
- community_16_contrib_legacy
- community_17_tests
- community_18_scripts_journal
- community_19_modules_backdoor
- community_1_cli_commands_utils
- community_20_poc_tui
- community_21_scripts_check_contract_manifest
- community_22_scripts_migrate_lazyown
- community_23_tools
- community_24_orphans
- community_2_cli
- community_3_modules_lazy_loop
- community_4_modules_lazyc2
- community_5_lazygui_panels
- community_6_static_js
- community_7_modules_lazyown_facts
- community_8_modules_opsec_scorer
- community_9_cli_commands_c2_profile_engine
- concepts
- queries
- REPORT
Agent Docs
- API
- API_p10
- API_p11
- API_p12
- API_p13
- API_p14
- API_p15
- API_p16
- API_p17
- API_p18
- API_p19
- API_p2
- API_p20
- API_p21
- API_p22
- API_p23
- API_p24
- API_p3
- API_p4
- API_p5
- API_p6
- API_p7
- API_p8
- API_p9
- ARCHITECTURE
- ARCHITECTURE_p2
- ARCHITECTURE_p3
- ARCHITECTURE_p4
- ARCHITECTURE_p5
- ARCHITECTURE_p6
- ARCHITECTURE_p7
- ARCHITECTURE_p8
- ARCHITECTURE_p9
- CONCEPTS
- GOTCHAS
- INDEX
- INDEX_p2
- INDEX_p3
- KB_backdoor
- KB_blueprints
- KB_cgi-bin
- KB_claude_md_orchestrator
- KB_cli
- KB_cli_p2
- KB_cli_p3
- KB_cli_p4
- KB_cli_p5
- KB_commands
- KB_commands_p2
- KB_commands_p3
- KB_commands_p4
- KB_config
- KB_contract
- KB_core
- KB_core_p2
- KB_devtools
- KB_extensions
- KB_hermes-lazyown
- KB_integrations
- KB_js
- KB_js_p2
- KB_js_p3
- KB_js_p4
- KB_lazyc2
- KB_lazygui
- KB_lazyown-docker
- KB_legacy
- KB_legacy_p2
- KB_misc
- KB_modules
- KB_modules_p10
- KB_modules_p11
- KB_modules_p12
- KB_modules_p13
- KB_modules_p14
- KB_modules_p15
- KB_modules_p16
- KB_modules_p2
- KB_modules_p3
- KB_modules_p4
- KB_modules_p5
- KB_modules_p6
- KB_modules_p7
- KB_modules_p8
- KB_modules_p9
- KB_modules_tools
- KB_palettes
- KB_panels
- KB_plugins
- KB_poc_tui
- KB_root
- KB_root_p2
- KB_rootkit
- KB_scripts
- KB_security
- KB_services
- KB_sinks
- KB_skills
- KB_skills_claude_md_orchestrator_tests
- KB_skills_p2
- KB_skills_p3
- KB_skills_p4
- KB_skills_p5
- KB_skills_tests
- KB_tests
- KB_tests_p10
- KB_tests_p11
- KB_tests_p12
- KB_tests_p13
- KB_tests_p14
- KB_tests_p15
- KB_tests_p16
- KB_tests_p17
- KB_tests_p18
- KB_tests_p19
- KB_tests_p2
- KB_tests_p3
- KB_tests_p4
- KB_tests_p5
- KB_tests_p6
- KB_tests_p7
- KB_tests_p8
- KB_tests_p9
- KB_theme
- KB_tools
- KB_widgets
- KB_win_rootkit
- KB_windows
- MEMORY
- SECURITY
- SYMBOLS
- SYMBOLS_p10
- SYMBOLS_p11
- SYMBOLS_p12
- SYMBOLS_p13
- SYMBOLS_p14
- SYMBOLS_p15
- SYMBOLS_p16
- SYMBOLS_p17
- SYMBOLS_p18
- SYMBOLS_p19
- SYMBOLS_p2
- SYMBOLS_p20
- SYMBOLS_p21
- SYMBOLS_p22
- SYMBOLS_p23
- SYMBOLS_p24
- SYMBOLS_p25
- SYMBOLS_p26
- SYMBOLS_p27
- SYMBOLS_p28
- SYMBOLS_p29
- SYMBOLS_p3
- SYMBOLS_p30
- SYMBOLS_p31
- SYMBOLS_p32
- SYMBOLS_p33
- SYMBOLS_p34
- SYMBOLS_p35
- SYMBOLS_p36
- SYMBOLS_p37
- SYMBOLS_p38
- SYMBOLS_p39
- SYMBOLS_p4
- SYMBOLS_p40
- SYMBOLS_p41
- SYMBOLS_p42
- SYMBOLS_p43
- SYMBOLS_p44
- SYMBOLS_p5
- SYMBOLS_p6
- SYMBOLS_p7
- SYMBOLS_p8
- SYMBOLS_p9
Recipes