发表机构
X32 Studio（X32工作室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出记忆-技能同构思想，将记忆作为技能载体，通过渐进式披露和受治理写入实现智能体复用，并给出部署系统中的治理与性能证据。
AI 中文摘要
记忆和技能在不改变权重的情况下提升智能体：记忆携带先验经验，技能携带可复用流程。将两者封装进存储、路由器、检索、反思和更新路径中，会使复用机制随积累而增长。这部分重复无需重建：技能（Skill）已经是蒸馏记忆的自然载体。在此，记忆组件即技能：常驻描述保存热线索，按需的this http URL保存较冷的索引和策展策略，reference/*.md文件保存详细记忆（层级L0 -> L1 -> L2）。受治理的1,024字符描述预算保持常驻索引的压缩性。记忆和能力因此共享一个渐进式披露载体；反思可演化任一技能。写入分叉：历史被追加，当前状态被重写并重新验证。在一个已部署系统中，最尖锐的识别是治理：存在4条写入条目，但只有1/4到达结算账本。在定价运行点，一条L1课程-1点增加1,313个首轮令牌，基线为1,462个令牌——在k=1时持平（1,313对1,365），在k=5时为4,949；会话总数仅相差1.18倍，末轮范围重叠；价格分解不可用。在选定的一个任务和一个模型上，暴露课程意味着更少的失败（有课程时为0/8或1/8，对比共享非并发的历史底线6/6，未对多重性进行调整）。该任务基于先前底线证据选择，因此这是选定任务的选择后存在性信号，而非确证性比率。常驻和BM25在两个小规模比较中未显示检测到的差异。这激发了一个候选RSI设计规则：一个技能载体家族，共享读取侧，受治理的独立写入。主体交付h:= P(D|A)在RQ1--RQ2中未进行仪器化；这评估了一个常驻索引实现。
英文摘要
Memory and skills improve agents without changing weights: memory carries prior experience, skills carry reusable procedures. Wrapping both in stores, routers, retrieval, reflection, and update paths makes reuse machinery grow with accumulation. Part of this duplication need not be rebuilt: a Skill is already a natural carrier for distilled memory. Here the memory component is a Skill: resident description holds hot cues, on-demand SKILL.md a colder index and curation policy, and reference/*.md files detailed memories (levels L0 -> L1 -> L2). A governed 1,024-character description budget keeps the resident index compressed. Memory and capability thus share one progressive-disclosure carrier; reflection evolves either Skill. Writes fork: history appended, current state rewritten and revalidated. In one deployed system, the sharpest identification is governance: 4 write entries exist, but only 1/4 reaches the settlement ledger. At the priced operating point, one L1 lesson-1 point adds 1,313 first-turn tokens against a 1,462-token baseline--tied at k=1 (1,313 versus 1,365), 4,949 at k=5; session totals differ by only 1.18x with overlapping last-turn ranges; price decomposition is unavailable. On one selected task with one model, exposing the lesson means fewer failures (0/8 or 1/8 with the lesson versus a shared non-concurrent 6/6 historical floor, unadjusted for multiplicity). The task was selected on prior floor evidence, so this is a selected-task post-selection existence signal, not a confirmatory rate. Resident and BM25 show no detected difference in two small comparisons. This motivates a candidate RSI design rule: one Skill carrier family, a shared read side, governed distinct writes. Body delivery h := P(D|A) is uninstrumented in RQ1--RQ2; this evaluates a resident-index implementation.
Comments18 pages, 2 figures
