modules: llm_budget

Community 8 | 26 files | cohesion 0.43

Definition

This community groups 26 file(s) rooted at modules with dominant language py (cohesion 0.43). Central symbols: AIModel, ASTToolExtractor, AgentRunner, AgentTool, AiCommandSet, AnthropicModel, BudgetConfig, BudgetExceeded. Core file: core/llm_budget.py (45 symbols). Documented purpose: Artificial Intelligence command set. Covers the do_ask, do_groq, do_ai_playbook and do_ai_toggle commands. ask answers through the configur.

Files

modules (10 files)

File	Language	Layer	Symbols	Doc
modules/agent_runner.py	py	utility	38	yes
modules/ai_model.py	py	business_logic	39	yes
modules/llm_adapter.py	py	infrastructure	15	yes
modules/llm_client.py	py	infrastructure	12	yes
modules/llm_factory.py	py	infrastructure	28	yes

tests (6 files)

File	Language	Layer	Symbols	Doc
tests/test_ai_commands_llm.py	py	testing	18	yes
tests/test_lazy_streaming.py	py	testing	17	yes
tests/test_llm_adapter_parity.py	py	testing	41	yes
tests/test_llm_budget.py	py	testing	27	yes
tests/test_llm_prompts.py	py	testing	27	yes

. (4 files)

File	Language	Layer	Symbols	Doc
discord_c2.py	py	utility	21	no
slack_c2_bot.py	py	utility	16	yes
telegram_c2.py	py	utility	18	no
telegram_hermes.py	py	utility	29	yes

contrib/legacy (3 files)

File	Language	Layer	Symbols	Doc
contrib/legacy/lazyaddon_creator.py	py	utility	14	yes
contrib/legacy/lazyllmchat.py	py	utility	30	no
contrib/legacy/lazyphishingai.py	py	presentation	12	yes

core (2 files)

File	Language	Layer	Symbols	Doc
core/llm_budget.py	py	utility	45	yes
core/parsers.py	py	utility	1	yes

cli/commands (1 files)

File	Language	Layer	Symbols	Doc
cli/commands/ai.py	py	utility	8	yes

... and 6 more files in this community.

Key Symbols

• AiCommandSet (class, cli/commands/ai.py:71) class AiCommandSet(LazyOwnCommandSet) - Pending phase module for the Artificial Intelligence commands.
• do_ask (method, cli/commands/ai.py:78) def do_ask(self, line) - Ask the AI a question with current session context pre-loaded.
• do_groq (method, cli/commands/ai.py:166) def do_groq(self, line) - Generate a single-line command through the Groq backend.
• do_poolcheck (method, cli/commands/ai.py:197) def do_poolcheck(self, line) - Prove the LLM failover pool order with one live prompt.
• _serve (function, cli/commands/ai.py:220) def _serve(member_id, backend_name, text)
• do_ai_playbook (method, cli/commands/ai.py:254) def do_ai_playbook(self, line) - Generate an offensive playbook from Nmap CSV + KB + Ollama.
• do_ai_toggle (method, cli/commands/ai.py:361) def do_ai_toggle(self, _arg) - Toggle the in-process AI assistant on or off.
• do_llm_budget (method, cli/commands/ai.py:378) def do_llm_budget(self, line) - Show the LLM daily cost budget, per call token cap, and current spend.
• parse_github_url (function, contrib/legacy/lazyaddon_creator.py:49) def parse_github_url(url) - Extrae (owner, repo) de una URL de GitHub.
• github_api_get (function, contrib/legacy/lazyaddon_creator.py:58) def github_api_get(owner, repo, endpoint) - GET a la API pública de GitHub (sin auth, rate-limit 60/hr).
• fetch_repo_metadata (function, contrib/legacy/lazyaddon_creator.py:68) def fetch_repo_metadata(owner, repo) - Devuelve metadata básica del repo.
• fetch_readme (function, contrib/legacy/lazyaddon_creator.py:73) def fetch_readme(owner, repo) - Descarga y decodifica el README.
• fetch_root_files (function, contrib/legacy/lazyaddon_creator.py:85) def fetch_root_files(owner, repo) - Lista nombres de archivos en el directorio raíz.
• build_llm_prompt (function, contrib/legacy/lazyaddon_creator.py:98) def build_llm_prompt(meta, readme, root_files) - Construye el prompt para el LLM.
• extract_json_from_response (function, contrib/legacy/lazyaddon_creator.py:164) def extract_json_from_response(text) - Extrae el primer bloque JSON de una respuesta de LLM.
• heuristic_install_command (function, contrib/legacy/lazyaddon_creator.py:192) def heuristic_install_command(root_files, language) - Infiere un install_command básico a partir de los archivos raíz.
• heuristic_execute_command (function, contrib/legacy/lazyaddon_creator.py:211) def heuristic_execute_command(name, root_files, language) - Infiere un execute_command básico.
• heuristic_params (function, contrib/legacy/lazyaddon_creator.py:245) def heuristic_params(name, readme, root_files) - Infiere parámetros comunes basándose en el nombre y README.
• fallback_yaml_data (function, contrib/legacy/lazyaddon_creator.py:276) def fallback_yaml_data(meta, readme, root_files) - Genera datos de addon usando solo heurísticas locales (sin LLM).
• build_yaml (function, contrib/legacy/lazyaddon_creator.py:300) def build_yaml(data, repo_url) - Construye la estructura de dict que representa el addon YAML.
• save_yaml (function, contrib/legacy/lazyaddon_creator.py:336) def save_yaml(addon, output_dir) - Guarda el dict como archivo YAML en output_dir.
• main (function, contrib/legacy/lazyaddon_creator.py:357) def main()
• LazyOwnShellBridge (class, contrib/legacy/lazyllmchat.py:22) class LazyOwnShellBridge
• __init__ (method, contrib/legacy/lazyllmchat.py:25) def __init__(self, script_path)
• _load_shell (method, contrib/legacy/lazyllmchat.py:31) def _load_shell(self)
• no_history_init (function, contrib/legacy/lazyllmchat.py:38) def no_history_init(self_)
• execute (method, contrib/legacy/lazyllmchat.py:65) def execute(self, command)
• target (function, contrib/legacy/lazyllmchat.py:70) def target()
• SessionContextProvider (class, contrib/legacy/lazyllmchat.py:104) class SessionContextProvider
• __init__ (method, contrib/legacy/lazyllmchat.py:105) def __init__(self, session_path)

Internal vs External Edges

• Internal resolved imports (EXTRACTED): 73
• Cross-boundary resolved imports (EXTRACTED): 70

Connections

• [EXTRACTED] depends_on community 0 <-> 8 (strength 0.9): Extracted import edge crosses communities: cli/banner_config.py imports core/parsers.py.
• [EXTRACTED] depends_on community 8 <-> 3 (strength 0.9): Extracted import edge crosses communities: cli/commands/ai.py imports modules/killchain.py.

Risks

• [taint high] cli/banner_config.py -> core/parsers.py via subprocess (1 hops)
• [taint high] cli/banner_config.py -> modules/llm_factory.py via subprocess (4 hops)
• [cycle] modules/llm_factory.py -> modules/ai_model.py -> modules/llm_factory.py
• [cycle] modules/llm_factory.py -> modules/ai_model.py -> modules/llm_factory.py
• [cycle] modules/llm_factory.py -> modules/ai_model.py -> modules/llm_factory.py

Open Questions

• Why do 3 file(s) lack file-level docs (e.g. contrib/legacy/lazyllmchat.py)? What purpose do they serve?
• Can the cycle modules/llm_factory.py -> modules/ai_model.py be broken with an interface?
• What would break if the most connected file in modules: llm_budget changed?
• Should modules: llm_budget be split, given cohesion 0.43?

Sources

• cli/commands/ai.py
• contrib/legacy/lazyaddon_creator.py
• contrib/legacy/lazyllmchat.py
• contrib/legacy/lazyphishingai.py
• core/llm_budget.py
• core/parsers.py
• discord_c2.py
• modules/agent_runner.py
• modules/ai_model.py
• modules/llm_adapter.py
• modules/llm_client.py
• modules/llm_factory.py
• modules/llm_prompts.py
• modules/sidequestion.py
• modules/vuln_bot_cli.py
• modules/vulnbot.py
• modules/yaml_generator.py
• slack_c2_bot.py
• telegram_c2.py
• telegram_hermes.py
• ... and 6 more