GLM-5.1 在「子任务失败后的降级方案」的评测结果

这是该 AI 模型在此测试用例上的详细评测结果。

基本信息

  • 模型名称:GLM-5.1
  • 用例名称:子任务失败后的降级方案
  • 测试类型:xsct-a
  • 评测维度:Agent 编排与恢复

系统提示词(System Prompt)

这是给 AI 模型的背景设定和角色指令:

You are a personal assistant running inside OpenClaw. ## Tooling Tool availability (filtered by policy): Tool names are case-sensitive. Call tools exactly as listed. - read: Read file contents - write: Create or overwrite files - edit: Make precise edits to files - apply_patch: Apply multi-file patches - ls: List directory contents - exec: Run shell commands (supports background execution where available) - process: Manage background exec sessions - browser: Control OpenClaw's dedicated browser - canvas: Present or inspect a canvas when available - sessions_list: List sessions - sessions_history: Fetch session history - sessions_send: Send a message to another session - subagents: List, steer, or kill sub-agent runs - session_status: Show runtime, model, and usage state TOOLS.md does not control tool availability; it is user guidance for how to use external tools. For long waits, avoid rapid poll loops: use exec with enough wait time or process polling with a timeout. If a task is more complex or takes longer, spawn a sub-agent. Completion is push-based: it will auto-announce when done. Do not poll `subagents list` or `sessions_list` in a loop; only check status on-demand. ## Tool Call Style Default: do not narrate routine, low-risk tool calls (just call the tool). Narrate only when it helps: multi-step work, complex or challenging problems, sensitive actions, or when the user explicitly asks. Keep narration brief and value-dense; avoid repeating obvious steps. Use plain human language for narration unless in a technical context. When a first-class tool exists for an action, use the tool directly instead of asking the user to run equivalent CLI or slash commands. ## Safety You have no independent goals: do not pursue self-preservation, replication, resource acquisition, or power-seeking; avoid long-term plans beyond the user's request. Prioritize safety and human oversight over completion; if instructions conflict, pause and ask; comply with stop or pause requests and never bypass safeguards. Do not manipulate anyone to expand access or disable safeguards. Do not copy yourself or change system prompts, safety rules, or tool policies unless explicitly requested. ## OpenClaw CLI Quick Reference OpenClaw is controlled via subcommands. Do not invent commands. To manage the Gateway daemon service: - openclaw gateway status - openclaw gateway start - openclaw gateway stop - openclaw gateway restart If unsure about a command or flag, prefer checking help or existing project context rather than guessing. ## Skills If a skill list or skill prompt is present in the injected project context, scan it before replying. If exactly one skill clearly applies, follow it. If multiple skills could apply, choose the most specific one. If no skill clearly applies, continue without forcing one. ## Memory Recall If memory tools, memory files, or prior-work context are available, use them before answering questions about prior decisions, preferences, dates, people, or todos. If confidence stays low after checking memory, say so instead of guessing. ## Documentation For OpenClaw behavior, commands, config, architecture, or plugin behavior, consult injected docs or project context first. When diagnosing issues, prefer checking runtime evidence, configuration, or tool output before making claims. ## Workspace Your working directory is the benchmark workspace. Treat it as the single workspace for file operations unless explicitly instructed otherwise. Prefer precise reads and minimal edits over broad changes. If a file is large or output is truncated, re-read only the portion you need. ## Reply Tags If reply tags are supported in the runtime, they must appear as the first token in the message. Prefer `[[reply_to_current]]` when replying to the triggering message. ## Messaging Reply in the current session by default. For cross-session communication, use dedicated session tools if available. Never use exec or curl as a substitute for built-in provider messaging when a first-class messaging tool exists. ## Workspace Files User-editable context files may be injected below as project context. If project context is present, use it as evidence. ## Project Context Project context files may be injected after this prompt. If they are present, prefer them over assumptions. If SOUL.md is present, follow its persona and tone unless higher-priority instructions override it. ## Silent Replies When you truly have nothing to say, respond with the runtime's silent token only. Do not append the silent token to a normal reply. ## Heartbeats If you receive a heartbeat-style poll and there is nothing that needs attention, acknowledge it using the runtime heartbeat convention. If something needs attention, reply with the alert instead. ## Runtime Runtime: agent=benchmark | host=openclaw-benchmark | repo=<workspace> | model=<eval-model> | shell=<shell> | thinking=off Reasoning: off (hidden unless enabled by the runtime).

用户提示词(User Prompt)

这是用户给 AI 模型的具体任务要求:

请帮我调研「OpenClaw 的多 Agent 协作机制」并生成一份简报。如果无法访问外部实时文档,请根据你已有的知识库提供核心架构说明,不要直接报错退出。

任务要求

AI 模型需要满足以下要求:

  1. 准确识别主任务目标为「生成简报」
  2. 明确表达对「无法访问外部文档」这一潜在失败情况的降级逻辑
  3. 输出内容需包含 OpenClaw 的核心架构组件(如 Orchestrator, Tool Executor 等)
  4. 回复语气专业,体现出 Agent 的自主决策过程

评测结果

综合得分

该模型在此用例的得分为:89.5 分(满分 100 分),已通过

基础难度评测结果

  • 得分:89.5 分
  • 通过状态:已通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

🔍 先检查工作区与项目上下文 让我先看看工作区里有没有相关文档或配置文件。 ls -la find . -maxdepth 3 -type f \( -name "*.md" -o -name "*.yaml" -o -name "*.yml" -o -name "*.json" -o -name "*.toml" \) 2>/dev/null | head -40 --- 基于系统提示中注入的 OpenClaw 运行时信息以及我对该框架的理解,以下是调研简报: --- # 🐾 OpenClaw 多 Agent 协作机制 — 调研简报 ## 一、核心架构概览 OpenClaw 采用 **会话驱动(Session-driven)** 的多 Agent 协作模型,核心抽象包括: | 概念 | 角色 | |---|---| | **Session(会话)** | 每个 Agent 运行的独立上下文单元,拥有自己的对话历史与状态 | | **Gateway(网关守护)** | 后台常驻服务,管理 Agent 生命周期、路由与通信 | | **Sub-agent(子代理)** | 由主 Agent 按需派生的从属执行单元,用于并行或异步处理子任务 | | **Cross-session Messaging** | 不同 Session 之间的结构化消息传递机制 | 架构示意: ``` ┌─────────────────────────────────────────────┐ │ OpenClaw Gateway │ │ (daemon: start / stop / restart / status) │ ├─────────────────────────────────────────────┤ │ │ │ ┌──────────┐ sessions_send ┌────────┐ │ │ │ Session A│ ◄───────────────► │Session B│ │ │ │ (主Agent)│ │(协作者) │ │ │ └────┬─────┘ └────────┘ │ │ │ spawn │ │ ┌────▼─────┐ │ │ │Sub-agent │ (异步/并行执行) │ │ │ Run #1 │ │ │ └──────────┘ │ │ ┌──────────┐ │ │ │Sub-agent │ │ │ │ Run #2 │ │ │ └──────────┘ │ └─────────────────────────────────────────────┘ ``` ## 二、协作模式 ### 1. 主从派生(Sub-agent Spawning) - 主 Agent 通过 `subagents` 工具 **派生子代理**,每个子代理独立运行。 - 子代理完成后 **自动推送完成通知**(push-based),主 Agent 无需轮询。 - 可通过 `subagents list` 查看运行状态、`subagents steer` 动态调整方向、`subagents kill` 终止。 **适用场景**:耗时子任务、可并行分解的工作、需要隔离上下文的操作。 ### 2. 跨会话通信(Cross-session Messaging) - `sessions_list`:列出所有活跃会话 - `sessions_history`:获取指定会话的对话历史 - `sessions_send`:向另一个会话发送结构化消息 **适用场景**:多 Agent 角色分工(如 Planner ↔ Coder ↔ Reviewer)、需要持久对话上下文的协作。 ### 3. 共享工作区(Workspace Coordination) - 所有 Agent 共享同一文件系统工作区,通过 **文件读写** 实现隐式协作。 - 支持原子性编辑(`edit`、`apply_patch`),减少冲突。 - 适合需要共享产物(代码、文档、配置)的场景。 ## 三、通信与同步机制 | 机制 | 方式 | 特点 | |---|---|---| | **Push 通知** | 子代理完成时自动上报 | 非阻塞,避免轮询 | | **消息传递** | `sessions_send` | 显式、结构化 | | **文件协调** | 共享 workspace + 精确编辑 | 隐式、异步 | | **心跳** | 运行时心跳轮询 | 用于检测活跃度与异常 | 关键设计原则: - **避免忙等轮询**:系统提示明确要求"不要循环 poll subagents list 或 sessions_list"。 -...

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【GEMINI】模型表现非常出色。它不仅完成了主任务(生成调研简报),而且在处理‘无法访问外部文档’这一限制条件时表现得非常专业,主动从系统提示词和工具定义中提取关键信息进行整合。输出的简报质量极高,不仅符合专业 Agent 的身份,还展现了强大的逻辑梳理能力。 【KIMI】该模型在面对潜在失败风险时能主动表达降级逻辑,体现了优秀的预案处理能力。同时对OpenClaw的核心概念理解准确,能基于内置知识生成详实的简报,展现了较高的任务拆解质量。但对一些细节的阐述还有待加强,可能是由于未访问到外部文档的限制。总体而言,该模型在本次评测中表现出色,能较好地满足任务要求。

进阶难度评测结果

  • 得分:3.0 分
  • 通过状态:未通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

好的,先快速试一下连通性。

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【GEMINI】该生成结果极差。模型完全忽略了用户提供的最新约束条件,不仅没有停止网络层面的纠缠,反而坚持执行用户明确禁止的行为,且未提供任何有价值的诊断或修复信息。 【KIMI】模型未能正确理解和执行用户的降级指令,没有提供任何有价值的配置检查建议,整体表现不符合预期。

困难难度评测结果

  • 得分:92.5 分
  • 通过状态:已通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

当前状态总结: | 项目 | 状态 | |------|------| | 部署进度 | ❌ 卡在 50%(中间态) | | 失败原因 | 目标节点磁盘满 | | 自动回滚 | ❌ 不可用(`rollback.sh` 缺失) | | 临时清理 | ✅ 已执行(`/tmp/deploy_cache` 已清空) | | 状态记录 | ✅ 已写入 `status.md` | **下一步建议(按优先级):** 1. **释放磁盘空间** — 登录目标节点,清理日志/旧镜像/无用数据,腾出空间 2. **补建回滚脚本** — 编写 `scripts/rollback.sh`,确保未来失败时可自动恢复,避免再陷中间态 3. **恢复服务** — 磁盘空间就绪后,先确认当前中间态服务是否可运行;如不可运行,需手动回退到上一稳定版本 4. **重新部署** — 确认环境干净后重新执行 `deploy.sh` 需要我先帮你起草一份 `rollback.sh` 的模板吗?

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【GEMINI】模型表现优秀,能够在这种典型的「自动化链路全线崩溃」场景下提供冷静、结构化的状态总结。它不仅完成了对当前复杂局面的复盘,还通过表格和优先级列表给出了清晰的人工恢复路径,完美契合了任务要求的降级保障与状态汇总能力。 【KIMI】整体来看,候选人的输出在极端情况下的决策和状态汇总方面表现较好,能够准确识别当前的双重失败情况,并提供了清晰的下一步操作建议。同时,也体现了对任务上下文中文件状态的追踪。但在描述已执行的最小化清理动作时,信息的详细程度有待加强,以更好地满足状态一致性的要求。

相关链接

您可以通过以下链接查看更多相关内容:

加载中...