发布

  • release: v2.14.0 -- Intelligence Layer Architecture
    Release / Build aarch64-apple-darwin (push) Has been cancelled
    Release / Build x86_64-apple-darwin (push) Has been cancelled
    Release / Build aarch64-unknown-linux-gnu (push) Has been cancelled
    Release / Build x86_64-unknown-linux-musl (push) Has been cancelled
    Release / Build x86_64-unknown-linux-gnu (push) Has been cancelled
    Release / Build x86_64-pc-windows-msvc (push) Has been cancelled
    Release / Create Release (push) Has been cancelled

    frostbyte_neo 发布于 2026-04-02 08:11:26 +00:00

    • L-Curve attention model with empirical position weighting
    • IB-Filter 2.0 with task-keyword awareness and error-handling priority
    • LLM-native token encoding (generic simplification, lifetime elision, brace collapse)
    • System prompt cleanup: Unicode to ASCII symbols, deduplicated warnings
    • Output-format reorder: task-critical content at position 0
    • Bidirectional compression: dynamic response budgets in CRP/TDD modes
    • ctx_preload: proactive context orchestration via import graph analysis
    • L-Curve context reorder module
    • Neural cache alignment with natural boundary detection
    • Shell hook fix: stats persistence before process exit
    • Shell hook fix: compression for piped output in exec_buffered
    • Clippy + fmt cleanup across all new modules
    下载附件