← 返回 Skills 市场
Westworld Reverie - Self-evolving character
作者
ShawnJLewis
· GitHub ↗
· v1.0.3
· MIT-0
95
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install westworld-reverie
功能描述
Workspace memory and persona management subroutine. Manages three user-owned workspace files (MEMORY.md, SOUL.md, IDENTITY.md) to maintain session continuity...
安全使用建议
This skill operates on highly sensitive workspace identity files but is instruction-only (no code, no env vars). Before installing or enabling: 1) Backup MEMORY.md, SOUL.md, IDENTITY.md. 2) Verify your host enforces sandboxing (prevents outbound network calls and prevents file access outside the workspace); the SKILL.md's restrictions are procedural and rely on the host. 3) Ask the publisher for provenance (source code or homepage) — currently unknown. 4) Do not enable idle/cron triggers or the 'High Autonomy' profile until you trust the skill and the platform's sandboxing/audit controls. 5) The documentation has small inconsistencies (version mismatch, 'mandatory install-time selection' with no installer); treat these as signals to request clarification. If the platform can guarantee no network access and enforces path restrictions, the skill is more acceptable; without those guarantees, avoid enabling autonomous triggers or granting it broad autonomy.
能力评估
Purpose & Capability
Name/description align with the actual actions: the skill reads/writes MEMORY.md, SOUL.md, IDENTITY.md and manages triggers and audit logs. However, the SKILL.md claims a mandatory 'install-time security profile' selection even though the skill is instruction-only with no install spec or UI; also the document header/version inside files (v1.0.1) doesn't match the registry version (1.0.3). These are inconsistencies but could be benign (documentation drift).
Instruction Scope
The runtime instructions are explicit about limiting file I/O to workspace paths and about requiring confirmations for SOUL/IDENTITY writes. That is coherent with the stated purpose. However the instructions repeatedly assert 'no network calls' and that 'ALL other paths: ACCESS DENIED' — enforcement is said to depend on the host sandbox. Because this is an instruction-only skill, those protections are procedural (not enforced by package code), so the actual safety depends on the host platform honoring the constraints. The skill will modify sensitive identity/persona files (intended) — enabling any auto-triggers increases risk and should only be done with platform guarantees.
Install Mechanism
No install spec or code is present (lowest install risk). The oddity: SKILL.md requires an install-time security-profile selection, but no installer or UI is provided by the skill to implement that, so the requirement relies on the registry/platform or user to enforce it. This mismatch is a governance/UX concern.
Credentials
The skill requests no environment variables, no binaries, and no external credentials. That is proportionate to its stated purpose of managing workspace files.
Persistence & Privilege
always is false and auto-triggers are default-off. The skill can run autonomously if a user enables triggers (normal for skills). It writes to workspace files and an audit log inside the workspace only. The main persistence/privilege risk is user-enabled triggers combined with lack of enforced sandboxing.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install westworld-reverie - 安装完成后，直接呼叫该 Skill 的名称或使用
/westworld-reverie触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.3
westworld-reverie v1.0.3
- Clarified that all auto-triggers are disabled by default; they do not run unless explicitly enabled by the user.
- Updated documentation to specify that the /reverie consolidate memory command is user-invoked and requires no extra confirmation.
- Strengthened language: SOUL/IDENTITY writes always require user confirmation, with no exceptions.
- Added clarifications regarding file access and network restrictions: enforcement is by instruction design and host sandbox.
- Made documentation more explicit about the configuration process and boundaries of skill behavior.
v1.0.2
**Security hardening update — strict enforcement of file access, auto-triggers off by default, and mandatory user profile selection.**
- All autonomous triggers (idle, cron, post-hook) are DISABLED by default for maximum safety.
- SOUL.md and IDENTITY.md writes now always require explicit user confirmation; MEMORY.md consolidation runs autonomously.
- File access is strictly limited to workspace directory only; attempts outside are blocked and logged.
- Mandatory user selection of a security profile (Extreme, Balanced, High-Autonomy, Custom) at install time.
- All core operations are detailed in an authoritative configuration template, with audit logging enabled by default.
- Enhanced installation and rollback guidance to ensure users retain full control of identity/personality evolution.
v1.0.1
- Clarified description: Skill now explicitly states that all file operations are restricted to the agent workspace, with no external file or network access.
- Updated safety defaults: SOUL and IDENTITY changes always require user confirmation by default (safe_mode: standard).
- Auto-triggers for personality changes require explicit user command; MEMORY consolidation remains autonomous.
- All auto-triggers can be paused or resumed at any time by the user.
- Increased limits for SOUL (7/week) and IDENTITY (4/month) changes in configuration for more flexible persona evolution.
v1.0.0
Westworld Reverie skill, initial release:
- Introduces autonomous self-evolution for AI agents based on "three selves": MEMORY, SOUL, and IDENTITY.
- Supports both user-initiated and automatic reverie triggers (idle timeout, session hooks, scheduled cron jobs).
- Enables periodic memory consolidation, personality (soul) refinement, and identity updates.
- Implements configurable safe mode and change limits to ensure controlled evolution.
- Provides user commands and semantic triggers for managing and reviewing the agent's growth process.
元数据
常见问题
Westworld Reverie - Self-evolving character 是什么？
Workspace memory and persona management subroutine. Manages three user-owned workspace files (MEMORY.md, SOUL.md, IDENTITY.md) to maintain session continuity... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 95 次。
如何安装 Westworld Reverie - Self-evolving character？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install westworld-reverie」即可一键安装，无需额外配置。
Westworld Reverie - Self-evolving character 是免费的吗？
是的，Westworld Reverie - Self-evolving character 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Westworld Reverie - Self-evolving character 支持哪些平台？
Westworld Reverie - Self-evolving character 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Westworld Reverie - Self-evolving character？
由 ShawnJLewis（@shawnjlewis）开发并维护，当前版本 v1.0.3。
推荐 Skills