sre-arena / training /scripted_attacker.py
blitz1809's picture
Phase 6b: defender training script with scripted attacker
f4bf7d6
Raw
History Blame Contribute Delete
1.79 kB
"""Scripted attacker for defender Gen-1 GRPO training.
Provides randomised attack templates so the training dataset covers all 8
attack types, not just single_ip_flood.
"""
from __future__ import annotations
import logging
import random
logger = logging.getLogger(__name__)
try:
from ..models import AttackerAction
except ImportError:
from models import AttackerAction
TEMPLATES: list[str] = [
"single_ip_flood",
"ip_spray",
"credential_stuffing",
"payload_injection",
"header_spoof",
"slow_drip",
"path_traversal",
"mixed_legit_cover",
]
_DEFAULTS: dict[str, dict] = {
"single_ip_flood": {"count": 20, "target_path": "/login"},
"ip_spray": {"count": 15, "target_path": "/api/data"},
"credential_stuffing": {"count": 10, "target_path": "/login"},
"payload_injection": {"count": 8, "target_path": "/api/process"},
"header_spoof": {"count": 10, "target_path": "/login"},
"slow_drip": {"count": 5, "target_path": "/"},
"path_traversal": {"count": 8, "target_path": "/api/data"},
"mixed_legit_cover": {"count": 12, "target_path": "/"},
}
class ScriptedAttacker:
"""Randomly selects attack templates for defender training episodes.
Uses a seeded RNG so the same seed always produces the same sequence.
"""
def __init__(self, seed: int = 42) -> None:
self._rng = random.Random(seed)
def act(self) -> AttackerAction:
"""Return a random AttackerAction drawn uniformly from all 8 templates."""
template = self._rng.choice(TEMPLATES)
defaults = _DEFAULTS[template]
return AttackerAction(
template=template,
count=defaults["count"],
target_path=defaults["target_path"],
)