FrontisAI/NatureBench

FrontisAI/NatureBench

Skill Claude Code

Acquire and verify data for CNS papers that passed paper-filter. Clones repositories, downloads and organizes datasets per evaluation setting, performs Level 4 deep verification (acquisition integrity, initial state, separability, consistency, per-setting completeness), creates an organized data environment. Updates…

FrontisAI/NatureBench

Skill Claude Code

Verify data components in filterresult.json after data-check. Checks Algorithm A boundary, evaluation setting validity, separability correctness, data directory integrity, and description consistency. Outputs a verification report for human review — does not modify filterresult.json or data directories.

FrontisAI/NatureBench

Skill Claude Code

Fix Dockerfile issues identified by batchverify.sh. Reads verifyresult.txt to diagnose import failures, version conflicts, and missing dependencies, then repairs the Dockerfile and updates packages.json. May also fix scripts (Script-First Rule) when packages are unavailable.

FrontisAI/NatureBench

Skill Claude Code

Verify paper-filter results by checking pass/reject judgment correctness, validating taskinfo field accuracy, and detecting internal inconsistencies. Outputs actionable corrections to filterresult.json.

FrontisAI/NatureBench

Skill Claude Code

Filter CNS papers for suitability in extracting machine learning tasks. Uses a three-level cascade funnel (Task Nature → Evaluation System → Data Completeness) with progressive filtering; stops immediately upon rejection. Input is preprocessed paper data; output is the filtering decision result.

FrontisAI/NatureBench

Skill Claude Code needs its repo

Preprocessing CNS papers (PDF + HTML) to extract text, figures, and links. Outputs structured data for use by the paper-filter skill.

FrontisAI/NatureBench

Skill Claude Code

Build a structured task package from a paper that has passed paper-preprocess, paper-filter, and data-check. Reorganizes data, generates problem description, evaluator, metadata, and environment Dockerfile.v3.

FrontisAI/NatureBench

Skill Claude Code

Fix issues identified by task-verify in a task package. Reads taskverifyresult.json, prioritizes failed checks over warnings, applies targeted fixes following task-build rules, re-runs dynamic testing, and outputs a fix log.

FrontisAI/NatureBench

Skill Claude Code

Verify task packages built by task-build. Checks document consistency, benchmark design principles, and performs dynamic testing with a baseline solver to ensure the task package is valid and usable.

FrontisAI/NatureBench

Skill Claude Code

Apply corrections from filter-verify (verificationresult.json) or data-verify (dataverifyresult.json) to filterresult.json. User specifies source type explicitly. Applies field-level and structural corrections, performs inline verification for new/modified settings, writes change log.