📦 scenique-prevision — 多智能体预测
v1.0.0基于 Menos 架构模拟多智能体集群（怀疑者、乐天派、战略家），结合 L1 物理账本与 L2 长期记忆层上下文，输出决策结果的概率预测报告，辅助高阶决策。
1· 200·0 当前·0 累计
by @shewingong下载技能包
最后更新
2026/3/11
安全扫描
OpenClaw
可疑
medium confidence该技能声称集成 Menos L1/L2 与 OpenClaw 网关，但包内缺失对应代码、配置或凭证，整体无法验证。
评估建议
This package looks like a stub: it promises Menos/L1-L2 and OpenClaw gateway integration but doesn't include the orchestrator code or any credentials/config to make those integrations work. Before installing or enabling it, ask the publisher for: (1) the missing orchestrator files (e.g., core_logic/prevision_swarm_v0.py) and an explanation of what they do; (2) exact configuration steps and what gateway endpoint/credentials it requires; (3) a data-flow explanation showing what context is read fro...详细分析 ▾
⚠ 用途与能力
The skill claims automatic Menos L1/L2 context pulls and OpenClaw subagent integration, yet there are no declared environment variables, endpoints, or included code to perform these integrations. The main Python file is a lightweight stub that only returns a reference to a non-included orchestrator file (core_logic/prevision_swarm_v0.py). This mismatch suggests the package is incomplete or misrepresented.
⚠ 指令范围
SKILL.md instructs use of the prevision_simulate tool and states it "pulls context from L1/L2" and "spawns" agents, but gives no concrete runtime steps, endpoints, or credentials. It also states a requirement for an OpenClaw gateway with subagent capabilities but does not specify how to configure or authenticate that gateway. The instructions are vague and grant broad implicit authority without the necessary operational details.
✓ 安装机制
No install spec is provided (instruction-only plus a small local Python file). Nothing is downloaded or written during install, so there is low immediate install risk. The presence of a single small stub script further suggests this is a placeholder rather than an installer that fetches additional code.
⚠ 凭证需求
SKILL.md requires an OpenClaw gateway but skill.json requires no environment variables and the package declares no primary credential. That is inconsistent: integration with an external gateway typically requires at least an endpoint and auth token/credentials, which are not declared or documented here.
✓ 持久化与权限
The skill is not 'always' enabled, is user-invocable, and doesn't request system-wide config paths or claim to modify other skills. There is no evidence of elevated persistence or privilege requests in the provided files.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv1.0.02026/3/11
- 首次发布 scenique-prevision 技能。 - 实现 Menos 架构的“Prevision”层。 - 模拟多智能体集群（怀疑者、乐天派、战略家）以预测决策结果。 - 集成 Physical Ledger（L1）与 Long-term Memory（L2）层获取上下文。 - 输出概率报告辅助决策。 - 需要支持子代理的 OpenClaw 网关。
● 无害
安装命令
点击复制官方npx clawhub@latest install scenique-prevision
镜像加速npx clawhub@latest install scenique-prevision --registry https://cn.longxiaskill.com