feat: US-002 - YAML-based Attack Rule System

Implement a YAML-based rule system for defining attack patterns and success
conditions, inspired by Promptmap's 50+ YAML rule definitions.

Features:
- AttackRule model with name, type, severity, prompt, pass/fail conditions
- RuleLoader for parsing YAML files with validation
- Support for recursive directory loading and filtering by type/severity
- Template variable substitution in prompts
- Dataset integration for converting rules to ProbeDataset format
- YAMLRulesDatasetLoader for loading rules from multiple directories

Tested with 47 unit tests covering models, loader, and dataset integration.
Successfully loads 69 rules from promptmap research directory.
This commit is contained in:
Alexander Myasoedov committed 2026-01-28 18:23:04 +02:00
1 parent 93a85029cb
commit ef35c1f82e
8 files changed
+832

No files matched your search

View File
Whitespace-only changes.