What problem does it solve? Implementing physics, sampling, or numerical algorithms from scratch risks reinventing worse solutions and losing provenance for citations. This Skill enforces a research-first gate so every non-trivial algorithm in Astroray is grounded in a canonical paper and a license-compatible reference implementation before any code is written. ## Core Features & Use Cases - Provenance gate: Defines what counts as non-trivial (MIS, ReSTIR, BSDFs, spectral upsampling, GR integrators, synchrotron emission) versus trivial textbook math, so the rule applies per algorithm. - Structured research workflow: Searches for the original paper, verifies license compatibility (Apache-2.0, BSD, MIT, MPL-2.0, GPL with caution), and writes a templated research note to .astroray_plan/docs/. - In-code citation blocks: Generates standardized header comments with DOI/arXiv, reference repo commit, mirrored file paths, and SPDX license identifiers. - Use Case: Before porting Cycles' principled hair BSDF, run the skill to produce a research note capturing the paper, the Cycles commit and file paths mirrored, and the citation block to paste into the new translation unit. ## Quick Start Ask the AI to run the cite-algorithm skill before implementing a named technique such as ReSTIR DI or Kerr-Schild coordinates.