将智能体技能转为可验证的图结构，提升执行可靠性与改进效率。
AIP: A Graph Representation for Learning and Governing Agent Skills
- 用有向图建模技能，节点为步骤，边为输入输出关系，支持精准编译。
- 编译后任务成功率从53%升至67%，平均奖励从0.60提至0.71（p=0.011）。
- 支持逐节点调试与重构，让技能优化变成可量化的迭代过程。
当前智能体技能多以自由文本形式存在，需在每次会话中重新解析和推导行为，导致实现复杂任务时可靠性下降，且技能创建与改进困难。为此提出代理指令协议（AIP），将技能建模为有向执行图：离散步骤作为节点，由确定性脚本或自然语言描述支撑，通过显式类型化输入/输出边连接，并由结构验证的YAML规范管理。一个编译元技能可将现有人工编写技能转换为该格式。实验显示，将技能转化为AIP后，Claude Sonnet在SkillsBench的27个真实任务上，平均任务奖励从0.60提升至0.71，通过率从53%增至67%（威尔科克森符号秩检验，p=0.011），赢下12项任务，仅败2项，13项平局，常耗时更少。图结构提供经验证的可运行单元，避免智能体重复推导代码与工具调用。此外，由于每个节点可独立验证、测试与定位，故障可精准诊断修复；两次人工编写的技能失败被定位到脚本层，调整AIP规范并重新编译后，两任务均零回退恢复（一项从0/5提升至5/5），使技能改进变为可度量的调优循环。该结构还支持语料级治理、技能自省，并自然适合作为强化学习中的动作空间。
原文摘要 · Abstract (English)
Agent Skills today consist largely of free-form prose requiring the agent to read, interpret, and re-derive how to act in every session. This imposes two compounding costs: reduced reliability on implementation-heavy tasks, and difficulty in skill creation and improvement, since editing prose is a fragile process that both humans and agents struggle with, particularly for domain-specific procedural knowledge underrepresented in model training. The Agent Instruction Protocol (AIP) addresses both by modeling a skill as a directed execution graph: discrete steps as nodes backed by deterministic scripts or natural-language descriptions, connected by explicit typed input/output edges, and governed by a schema-validated YAML specification. A compiler meta-skill translates existing human-written skills into this form. The benefits are twofold. First, compiling human-written skills to AIP raised Claude Sonnet's mean task reward from 0.60 to 0.71 and pass rate from 53% to 67% across 27 real agent tasks from SkillsBench - a statistically significant gain (Wilcoxon signed-rank p = 0.011), winning 12 tasks to 2 with 13 ties - often in less wall-clock time. The graph delivers vetted, runnable units to the agent rather than asking it to re-derive code, commands, and tool calls from natural language. Second, on creation and improvement, because each skill is schema-validated, functionally testable, and addressable node-by-node, failures can be diagnosed and repaired precisely. Two authored-skill failures were traced to the script level. After adjusting the AIP spec and recompiling, both recovered with zero regressions (one task going from 0/5 to 5/5), turning skill improvement into a measurable tuning loop rather than a prose rewrite. That same graph structure supports corpus-level governance and skill introspection, and provides a natural action space for reinforcement learning over skills.
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。