Paper Deep Reader

**Expanded support and structure for diverse research paper types.** - Added detailed adapters for benchmark, dataset, replication, and synthesis papers, broadening supported paper types. - Introduced more targeted evidence checklists (e.g., for ablation, reproducibility, proof rigor, benchmarking fairness). - Updated routing and execution protocol: require explicit selection of one primary adapter plus up to three relevant evidence packs per paper. - Enhanced instructions for reading, mapping, and routing papers, with emphasis on identifying main intellectual load and failure risks. - Rewrote core rules for clarity, including process for cases where papers are surveys, syntheses, or have complex contributions. - Removed obsolete ML-specific adapters and checklists, replacing them with finer-grained alternatives.

**Major update introducing detailed adapters, checklists, and structured protocols for paper deep reading.** - Added domain-specific adapters and checklists for theory, ML, empirical economics, physics, quantitative finance, systems, and more. - Replaced the basic note template with a new `note-template-base.md` plus explicit output contract and detailed workflow. - Introduced required internal structures: paper map, notation table, claim-evidence matrix, and limitation ledger. - Provided a suite of structured scripts to aid first-draft creation and reduce drift from the paper. - Clarified purpose, execution sequence, and mandatory separation of authors' claims vs. reader evaluation. - Updated documentation to guide precise, evidence-based technical reading and note writing across disciplines.

- Initial release of paper-deep-reader skill. - Reads a selected research paper thoroughly and produces a detailed markdown note with equations, evidence, limitations, and implementation details. - Enforces clear separation between the paper’s claims and the reader’s evaluation. - Output is pedagogical, precise, and tailored to the type of paper (theory, empirical, systems, etc.). - Adheres to a strict workflow and output standard for graduate-level clarity. - Supports flexible saving: saves notes as `detailed-note.md` next to the paper by default or to a specified directory/filename.