Project Shimmer is a constitutional multi-agent system for reviewing regulated documents. Fixed-role agents work in stages on an append-only message bus, under written seed laws: producers and auditors are always different model families, so no model checks its own work. A run returns flagged provisions, cited findings and proposed amendments, each grounded in an operator-defined convention and a precise reference into the source corpus. It is multilingual, including right-to-left scripts, and domain-agnostic by construction.
SILICA-Bench measures whether a frontier model can audit the causal integrity of an observational research design: whether it catches the silent failures a trained methodologist would catch, such as a weak instrument, an absent diagnostic, or a method asserted outside its regime. Two decisions carry the design. Ground truth exists by construction, because known violations are injected into synthetic designs before the model sees anything. And the setup is de-primed: the model receives raw, unlabeled material and no checklist, because a checklist measures recall, not judgment. On its first run the system recovered a deliberately flawed design and flagged a planted weak instrument from raw data.
Detection Dance is an adversarial evaluation study of a commercial AI-text detector, probing what the detector actually responds to rather than what it claims to measure.
Values Pipeline, the LSE dissertation, classified approximately 25,700 UN General Debate sentences from 61 countries against ten World Values Survey dimensions, using SBERT embeddings and CatBoost classifiers with human-validated labels, a permutation-tested divergence index, and an explicit error-propagation model. Awarded a distinction; a journal article is in preparation.
Cosmic Horror Index: an NLP pipeline scoring cosmological traditions across ten weighted axes. A replication study that stress-tested a published instrumental-variable paper and showed its instrument weakening under unreported transformations. A survey-attrition classifier built on gradient boosting and neural networks. The rest is on GitHub.
SILICA-Bench and Detection Dance have no public repository. A detailed walkthrough of either, design, method and results, is available on request.