skill_context_cache

Context caching preserves continuity across pipeline steps by storing design decisions, measurements, and cross-references in persistent storage. Retrieve cached data in later stages to propagate decisions and maintain traceable provenance across spec, analysis, implementation, and benchmarking phases.

Yes. Cross-referencing links spec decisions to analysis results and benchmarks, creating traceable connections. Validation ensures references stay consistent and up-to-date across all stages, supporting reproducible research workflows.

Validation operations detect inconsistent or stale references across linked artifacts. The Skill maintains cross-reference integrity, flagging broken links and ensuring cached data remains current throughout multi-stage workflows.

Store operations persist decisions to .cache/context.yaml with linked metadata. Retrieve operations load cached decisions in downstream stages, enabling reuse without re-computation and establishing decision provenance across the pipeline.

Yes. It's designed for research pipelines requiring cross-stage memory—spec, analysis, implementation, and benchmarking phases all access the same cached context, eliminating information loss between stages and enabling reproducible, auditable workflows.