← 返回 Skills 市场
Fast Response Optimizer
作者
jason-tiger
· GitHub ↗
· v1.0.0
· MIT-0
148
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install fast-response-optimizer
功能描述
Response speed optimizer - implements reply-first-then-process, parallel tool calls, and memory file caching
使用说明 (SKILL.md)
Fast Response Optimizer ⚡
Reduce response time from 5-20 seconds to under 1 second
Core Optimization Strategies
1. Reply First, Process Later
Rule:
- Receive request → Immediately reply "Received/Processing" → Execute in background → Push result when done
Implementation:
// Quick reply
await message.send("Received, processing...");
// Background execution
const result = await processInBackground(task);
// Push result
await message.send(result);
2. Parallel Tool Execution
Rule:
- Execute independent tool calls in parallel instead of waiting sequentially
Implementation:
// Before (sequential)
const result1 = await toolA();
const result2 = await toolB();
const result3 = await toolC();
// After (parallel)
const [result1, result2, result3] = await Promise.all([
toolA(),
toolB(),
toolC()
]);
3. Memory File Caching
Rule:
- Load cache every 1 minute via scheduled task
- Reload cache if more than 1 minute since last message
- Cache stored in SESSION-STATE.md
Implementation:
// Cache structure
const cache = {
lastUpdate: timestamp,
userProfile: {...},
memorySummary: {...},
skills: [...]
};
// Check if refresh needed
if (now - lastMessage > 60000 || now - cache.lastUpdate > 60000) {
await refreshCache();
}
Trigger Conditions
| Scenario | Trigger Method |
|---|---|
| Auto-trigger | Execute on every message processing |
| Scheduled refresh | Cron job every 1 minute |
| Manual trigger | User inputs "refresh cache" |
Cache Files
SESSION-STATE.md # Active state + cache
├── cache/ # Cache directory
│ ├── user_profile.json # User profile cache
│ ├── memory_summary.json # Memory summary cache
│ └── skills_list.json # Skills list cache
└── last_refresh.txt # Last refresh timestamp
Performance Goals
| Metric | Before | After |
|---|---|---|
| First Response | 5-20s | \x3C1s |
| Tool Calls | Sequential accumulation | Parallel execution |
| File Reading | Read every time | Cache hit |
| Cache Refresh | None | Every 1 min / >1 min |
Usage
# Initialize optimizer
node skills/fast-response-optimizer/bootstrap.js
# Manual cache refresh
node skills/fast-response-optimizer/scripts/cache-manager.js refresh
# Check cache status
node skills/fast-response-optimizer/scripts/cache-manager.js status
# Update message timestamp
node skills/fast-response-optimizer/scripts/cache-manager.js message
Make every response lightning fast ⚡
安全使用建议
This skill implements caching and parallel execution as advertised, but it also reads many workspace files (SOUL.md, USER.md, AGENTS.md, SESSION-STATE.md, etc.) and exposes a helper that runs arbitrary shell commands. Before installing: (1) inspect the workspace files named in scripts to ensure they contain no secrets or tokens; (2) consider running the skill in an isolated environment (sandbox/container) if you will use it on a workspace with sensitive data; (3) if you don't need shell execution, remove or disable parallelExecCommands or restrict the code so it can't run arbitrary commands; (4) trust the source (github.com/opendolph) and review the code yourself or have a developer audit it. If you cannot perform these checks, treat the skill as risky and avoid installing on production environments.
能力评估
Purpose & Capability
The name/description (reply-first, parallel calls, memory caching) aligns with the code: it implements cache loading, parallel execution helpers, and a bootstrap hook. However the parallel utilities include a function that runs arbitrary shell commands (parallelExecCommands using child_process.exec), which is not mentioned in the SKILL.md and is a more powerful capability than the description implies.
Instruction Scope
SKILL.md instructs auto-trigger on every message, scheduled 1-minute refresh, and caching of workspace memory files. The implementation reads multiple workspace files (SOUL.md, USER.md, MEMORY.md, AGENTS.md, SESSION-STATE.md, HEARTBEAT.md, WORKING.md) and writes cache files under a cache/ directory. That file I/O is consistent with 'memory file caching' but it gives the skill access to potentially sensitive workspace files. The SKILL.md does not call out the ability to run shell commands, but the code exposes that capability.
Install Mechanism
No install spec; files are included in the package. No remote downloads or installers are present, so there is no additional install-time network risk.
Credentials
The skill requests no environment variables, but it reads arbitrary workspace files and writes cache metadata. Reading AGENTS.md, SESSION-STATE.md and other workspace files can expose tokens, config, or secrets if they exist. The ability to execute shell commands (parallelExecCommands) expands the credential/privilege risk because a compromised or misused skill could run commands that read or transmit secrets.
Persistence & Privilege
The skill is not marked always:true and does not modify other skills or system-wide settings. It does update its own '.last-message' timestamp at bootstrap and writes cache files in the workspace, which is expected for a caching tool.
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 fast-response-optimizer - 安装完成后,直接呼叫该 Skill 的名称或使用
/fast-response-optimizer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
v1.0.0
- Initial release of fast-response-optimizer.
- Implements "reply-first-then-process" for sub-second acknowledgment of incoming requests.
- Supports parallel execution of independent tool calls to reduce processing time.
- Adds in-memory file caching with scheduled (every 1 minute), message-triggered, and manual refresh options.
- Stores active cache in SESSION-STATE.md and related files for efficient state handling.
- Provides CLI scripts for cache refresh, status, and message timestamp management.
元数据
常见问题
Fast Response Optimizer 是什么?
Response speed optimizer - implements reply-first-then-process, parallel tool calls, and memory file caching. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 148 次。
如何安装 Fast Response Optimizer?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install fast-response-optimizer」即可一键安装,无需额外配置。
Fast Response Optimizer 是免费的吗?
是的,Fast Response Optimizer 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Fast Response Optimizer 支持哪些平台?
Fast Response Optimizer 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 Fast Response Optimizer?
由 jason-tiger(@opendolph)开发并维护,当前版本 v1.0.0。
推荐 Skills