解决异构编程智能体中技能运行时状态不准确的问题。
Evidence-Calibrated Runtime Reconstruction for Agent Skills Across Heterogeneous Coding Agents
- 通过四类证据等级重建技能生命周期阶段，区分真实与未知状态。
- 在126次执行中精确匹配源会话，识别出三种不同语义的异常行为。
- 适合需要高可信度调试与可观测性的智能体系统开发者使用。
Agent Skills 将可复用的指令和资源封装为工具调用型语言模型代理的技能。渐进式加载导致会话、模型或工具中心的追踪难以准确反映失败边界：技能可能被发现但无法激活，激活后无说明，或看似成功却无独立验证结果。我们提出技能运行时智能（Skill Runtime Intelligence），一种被动运行时智能系统，可在异构环境中重建支持的技能生命周期阶段，同时将不支持的阶段保留为未知。其运行全景图（Run Panorama）以四类证据等级分离不可变事件、确定性关系、推断诊断与受控结果；支持可选追踪导入及 OTLP/HTTP 导出，兼容现有可观测性部署。在六个冻结仓库配置、三个编码代理及七种清洁或故障注入条件下，全部126次执行均保持源工作树不变，且每项对应唯一源会话。然而适配器揭示三种不同语义：无技能运行；完整运行但无类似失败事件；或在每个操作失败与清洁会话中均出现类似失败事件。在七模板诊断研究中，语义别名与全景图定位相同六处非清洁边界，但在具体行为与状态上存在差异；两者原始视图在全部18个清洁案例中均报告失败状态，而全景图则无一报告。已知规则图符合126/126个冻结契约，而第二个模型仅完成378次调用中的228次。这些观察推动了可执行适配器的资格评估，并表明事件存在性不等于边界保真度，复合精确度分数会掩盖不同错误，模型解释不应覆盖确定性事实。
原文摘要 · Abstract (English)
Agent Skills package reusable instructions and assets for tool-using language-model agents. Progressive loading creates failure boundaries poorly represented by session-, model-, or tool-centric traces: a Skill can be discovered but not activated, activated without instructions, or appear successful without an independently verified outcome. We present Skill Runtime Intelligence, a passive runtime-intelligence system that reconstructs supported Skill-lifecycle stages across heterogeneous harnesses while preserving unsupported stages as unknown. Its Run Panorama separates immutable events, deterministic relations, inferred diagnoses, and controlled outcomes with four evidence grades; optional trace import and OTLP/HTTP export support existing observability deployments. Across six frozen repository profiles, three coding agents, and seven clean or fault-injected conditions, all 126 executions preserve source worktrees and each correlates to exactly one source session. Yet adapters expose three distinct semantics: no Skill runs; complete runs but no failure-like events; or failure-like events in every operational-failure and clean session. In a seven-template diagnostic study, semantic aliases and Panorama localize the same six non-clean boundaries but differ in exact/status behavior; both Raw views emit a failure status on all 18 clean cases, while Panorama emits none. A known-rule graph conforms to 126/126 frozen contracts, whereas a second model completes only 228/378 calls. These observations motivate executable adapter qualification and show that event presence is not boundary fidelity, composite exact scores mask distinct errors, and model explanations must not overwrite deterministic facts.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。