Loading repository dataβ¦
Loading repository dataβ¦
AkshatRaj00 / repository
π§ Mythos AI Core β Autonomous Cognitive Agent Framework Mythos AI Core is a modular autonomous AI agent framework designed for intelligent system orchestration, real-time telemetry analysis, multi-model reasoning, defensive security monitoring, and adaptive execution workflows. Built with a scalable architecture inspired by next-generation AI
A transparent discovery signal based on current public GitHub metadata.
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
A modular, production-ready AI cognitive agent runtime built in Python β featuring priority-based orchestration, TTL-aware memory, sandboxed execution, structured telemetry, and a pluggable recon pipeline.
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β cognitive-agent-core β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ€
β MasterOrchestrator β RuntimeExecutor β SandboxActuator β
β β β β β
β AgentEngine CognitiveBrain MemoryStore (TTL) β
β β β β
β CyberScout Telemetry β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
| Module | What it does |
|---|---|
MasterOrchestrator | Priority-based task queue with FIFO tie-breaking and lifecycle hooks |
RuntimeExecutor | Retry logic with configurable exponential backoff |
SandboxActuator | Wall-clock timeout + POSIX memory-limit enforcement |
MemoryStore | Thread-safe key-value store with per-entry TTL and background pruning |
CognitiveBrain | LLM reasoning layer (Gemini-backed) |
AgentEngine | Main entry point wiring all modules together |
CyberScout | Pluggable recon pipeline with TTL result caching |
Telemetry | Structured event tracing with ring-buffer and summary stats |
AgentConfig | Env-aware config dataclass with field validation |
git clone https://github.com/AkshatRaj00/cognitive-agent-core.git
cd cognitive-agent-core
python -m venv .venv && source .venv/bin/activate
pip install -e .[dev]
export GEMINI_API_KEY="your-key-here"
python agent_engine.py
pytest -v --cov=. --cov-report=term-missing
from memory_store import MemoryStore
store = MemoryStore(default_ttl=3600) # 1-hour default
store.set("session:abc", {"user": "akshat"}, ttl=300) # 5-min override
print(store.get("session:abc")) # {'user': 'akshat'}
from master_orchestrator import MasterOrchestrator
orch = MasterOrchestrator(
on_task_success=lambda name, result: print(f"{name} done: {result}")
)
orch.enqueue("fetch_data", lambda: "data", priority=1) # highest
orch.enqueue("log_event", lambda: "logged", priority=10) # lowest
orch.run_all() # fetch_data runs first
from runtime_executor import RuntimeExecutor
executor = RuntimeExecutor(max_retries=3, backoff_base=2.0)
result = executor.execute(my_api_call, task_name="gemini.generate")
All settings are configurable via environment variables (prefix: AGENT_) or programmatically:
export AGENT_MODEL_NAME=gemini-1.5-flash
export AGENT_TEMPERATURE=0.5
export AGENT_MAX_RETRIES=5
export AGENT_DEBUG=true
MIT Β© Akshat Raj