v0.1.0 — first public release
First public release of the Narrative Harm Classifier: a rule-based narrative harm classification engine with escalation-chain tracking, a templated benchmark suite, multilingual support (English/Spanish/French/Russian/Arabic verified, Igbo/Yoruba/Hausa experimental), a curated dog-whistle lexicon, counter-narrative guidance, and tamper-evident provenance hashing.
Install: pip install narrative-harm-classifier
See the README for the full feature set, architecture, and known limitations.