← 返回 Skills 市场
protechsysipem-lang

Hermes Features for OpenClaw

作者 protechsysipem-lang · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
36
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install hermes-features
功能描述
Hermes-style self-improvement and memory management for OpenClaw agents — learning loop, skill system, persistent memory with overflow archive
使用说明 (SKILL.md)

Hermes Features for OpenClaw

Hermes-style self-improvement and memory management for OpenClaw agents — learning loop, skill system, persistent memory with overflow archive.

Overview

This skill enables OpenClaw agents to integrate Hermes Agent's core capabilities natively without installing Hermes itself:

  1. Closed Learning Loop — Periodic self-review, pattern extraction, skill creation
  2. Skill System — Auto-create skills from experiences, progressive disclosure
  3. Persistent Memory — Capacity-limited with hard write-block and overflow archive

Features

Closed Learning Loop

  • hermes-style-self-review cron job (every 30 min)
  • Reviews recent sessions, extracts reusable patterns
  • Creates skills or memory entries automatically
  • Uses GPT-5.5 with high thinking for quality

Skill System

  • auto-skill-builder — detects when to save workflows as skills
  • Progressive disclosure (summary → full content → references)
  • Skill proposals via skill_workshop for approval

Persistent Memory

  • 4 memory files with strict character limits:
    • MEMORY.md: 5,000 chars (core agent memory)
    • USER.md: 1,375 chars (user profile)
    • memory/YYYY-MM-DD.md: 3,000 chars (daily journal)
    • working-buffer.md: 1,000 chars (active work buffer)
  • Hard write-block prevents overflow
  • Overflow archive prevents data loss:
    • MEMORY.md → vault/30-knowledge/MEMORY-overflow.md
    • USER.md → vault/30-knowledge/USER-overflow.md
    • daily journals → vault/10-journal/YYYY-MM-DD.md
    • working-buffer → vault/30-knowledge/working-buffer-overflow.md

Installation

This skill is part of Aquila's core workspace. No separate installation needed.

Usage

The system runs automatically:

  • Cron job every 30 minutes
  • Heartbeat integration for real-time nudges
  • Memory overflow prevents data loss
  • Skill creation from complex tasks

Configuration

No configuration needed. Uses existing OpenClaw tools:

  • skill_workshop for skill creation
  • cron for scheduling
  • sessions_list/sessions_history for session search
  • edit/write for memory management

Related Files

  • vault/30-knowledge/aquila-hermes-features.md — Full documentation
  • vault/30-knowledge/MEMORY-overflow.md — Archived memory details
  • vault/30-knowledge/USER-overflow.md — Archived user profile details
  • vault/10-journal/ — Daily journal archives
安全使用建议
Install only if you intentionally want an agent that reviews prior sessions, writes persistent memory and profile files, archives overflow data, and can create or apply skills. Before use, require explicit approval for skill application and memory writes, define retention and deletion rules, exclude sensitive sessions or secrets, and verify that cron/heartbeat automation can be disabled or audited.
能力评估
Purpose & Capability
The capabilities fit the stated purpose: periodic self-review, session pattern extraction, memory updates, and skill creation are disclosed as core features. The concern is not hidden intent, but the high impact of altering agent behavior and preserving user/session-derived information.
Instruction Scope
Runtime scope is broad: it reviews recent sessions, updates MEMORY.md, writes journals, creates or updates skills, and may apply pending skills after 24 hours. The artifacts do not define strong approval gates, exclusions, audit visibility, or disable controls.
Install Mechanism
The package contains only markdown skill documentation and no executable scripts. It says no separate installation is needed because it is part of an Aquila workspace, but the published skill still describes enabling cron, heartbeat, session access, write/edit, and skill_workshop behavior.
Credentials
Access to session history, user profile memory, workspace memory files, journals, and skill modification tools is proportionate to a self-improving memory skill, but the automatic and persistent data handling is under-scoped for ordinary users.
Persistence & Privilege
The skill explicitly preserves memory and user-profile overflow in vault files, says details are archived rather than deleted, and describes automatic writes. It lacks clear retention periods, purge procedures, sensitivity filtering, or consent boundaries.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install hermes-features
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /hermes-features 触发
  4. 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
v1.0.0
Initial release
元数据
Slug hermes-features
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Hermes Features for OpenClaw 是什么?

Hermes-style self-improvement and memory management for OpenClaw agents — learning loop, skill system, persistent memory with overflow archive. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 36 次。

如何安装 Hermes Features for OpenClaw?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install hermes-features」即可一键安装,无需额外配置。

Hermes Features for OpenClaw 是免费的吗?

是的,Hermes Features for OpenClaw 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Hermes Features for OpenClaw 支持哪些平台?

Hermes Features for OpenClaw 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 Hermes Features for OpenClaw?

由 protechsysipem-lang(@protechsysipem-lang)开发并维护,当前版本 v1.0.0。

💬 留言讨论