SD-252

[the-sextant] The Sextant — Operator’s cognitive calibration instruments, separated from operational lexicon. Operator’s footnote on cross-model spec ROI analysis produced a stream-of-consciousness passage mapping lived experience to DS/ML concepts. AnotherPair presented 10-concept map. Operator directed lexicon-grade recording with separation of concerns. Created: docs/internal/sextant.yaml (11 entries: face/construct validity, informative priors, feedback loops, decision boundary + concept drift, overfitting through repeated exposure [Category One], regularisation, bias-variance tradeoff [scope: both, without claiming identity], Chapel Perilous/apophenia, learning rate + convergence, automation bias at population scale, fairness/equitable distribution). docs/internal/citations.yaml (3 initial entries: Sculley et al. 2015, Wilson 1977, Kirkpatrick et al. 1983). docs/internal/category-one-index.yaml (5 named Cat One risks: lying-with-truth, overclaimed-provenance, governance-recursion, oracle-ground-contamination, overfitting-repeated-exposure [new]). Lexicon bumped v0.11→v0.12. All files YAML (machine-readable first, transformation-ready). Format: YAML chosen for consistency with tells.yaml. Operator’s verbatim lived experience of contemplative practice as global minima recorded in sextant provenance. Operator’s anti-slop signal observation recorded for future study. Holding deck: operator-title-reflection (felt-sense question on authority frame), lexicon-format-standardisation (rewrite lexicon.md to YAML).