← 返回 Skills 市场
Clawzempic
作者
anna-claudette
· GitHub ↗
· v1.0.0
581
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install clawzembic
功能描述
Lighthouse-style efficiency audit for OpenClaw. Scores your instance A+ to F across 6 categories (context injection, cron health, session bloat, config, skil...
使用说明 (SKILL.md)
Clawzembic — Weight Loss for Your OpenClaw Instance
Lighthouse-style audit for OpenClaw efficiency. Run it, get a grade (A+ to F), shed the bloat.
Scans your installation and scores it across six critical categories: context injection, cron health, session bloat, config health, skill bloat, and transcript size. You get one overall letter grade plus category-by-category breakdown with actionable findings.
Zero dependencies. Python 3.8+ stdlib only.
Quick Start
# Audit this machine
bash skills/clawzembic/lean-audit.sh
# Audit a remote instance (VM, etc.)
bash skills/clawzembic/lean-audit.sh --remote user@host
# JSON output for dashboards/integrations
bash skills/clawzembic/lean-audit.sh --json
# Show automated fix suggestions
bash skills/clawzembic/lean-audit.sh --fix
# Custom .openclaw directory
bash skills/clawzembic/lean-audit.sh --dir /path/to/.openclaw
What It Checks
| Category | Weight | What Gets Flagged |
|---|---|---|
| Context Injection | 25% | Oversized MEMORY.md, bloated workspace files eating tokens |
| Cron Health | 25% | Wrong models, high thinking on routine tasks, main-session pollution |
| Session Bloat | 20% | Stale sessions hogging context, waste ratio above 35% |
| Config Health | 15% | Heartbeat too frequent, subagent defaults, missing compaction |
| Skill Bloat | 10% | Too many skills injected, unused skills inflating system prompt |
| Transcript Size | 5% | Disk hogs, oversized transcripts (>10MB files) |
Grade Scale
- A+/A (90-100): Lean machine 💪
- B+/B (75-89): Good shape, minor tweaks
- C+/C (60-74): Needs a diet 🍕
- D+/D (45-59): Significant bloat 🫠
- F (\x3C45): Emergency Clawzembic needed 💊
Agent Usage
When Jeffrey asks you to audit or optimize an OpenClaw instance:
- Run
bash skills/clawzembic/lean-audit.sh(or--remote user@claudettefor the VM) - Review the report and present key findings with context
- Offer
--fixmode for automated remediation if score \x3C 75
For remote instances, ensure SSH key-based auth is configured. The skill uses SSH to execute the audit remotely — no agent installation needed on the target.
安全使用建议
This skill appears to do what it claims: auditing OpenClaw directories and optionally running on a remote host via SSH. Before installing or running it on production data: 1) Review the remainder of lean-report.py (the audit logic beyond the truncation) to confirm there are no network calls, hidden endpoints, or code that exfiltrates files. 2) Test in a staging environment and run with --json or a dry-run to inspect outputs before using --fix. 3) If you use --remote, only SSH into hosts you control and understand that the script will run with the remote user's privileges. 4) Back up important transcripts/config before applying any automated fixes. If you want higher assurance, ask the publisher for a provenance/source URL or sign-off; the README references ClawHub but the skill metadata/homepage is missing.
功能分析
Type: OpenClaw Skill
Name: clawzembic
Version: 1.0.0
The skill is classified as suspicious due to its use of SSH for remote execution and the `--dir` argument allowing arbitrary path specification for auditing. While these capabilities are explicitly documented and intended for legitimate auditing purposes, they represent powerful functions that, if misused (e.g., via prompt injection against the agent to provide a malicious remote host or directory path), could lead to unauthorized remote command execution or arbitrary file access. The `lean-audit.sh` script quotes the `$REMOTE` variable in the `ssh` command, mitigating direct shell injection into the `ssh` command itself, and the `lean-report.py` script is read-only, only printing fix suggestions rather than executing them. However, the inherent power of these operations warrants a 'suspicious' classification as a vulnerability risk, not as intentional malware.
能力评估
Purpose & Capability
Name/description match the implementation: scripts scan .openclaw workspace, agents, and cron stores to produce an efficiency score. Required binary (python3) and declared zero-deps are consistent with the Python stdlib implementation.
Instruction Scope
SKILL.md and the scripts instruct the agent to read files under the OpenClaw install (workspace files, agent session stores, cron job JSON, transcripts). That is necessary for an audit, but these files can contain sensitive data (transcripts, session content). The remote mode streams the audit script over SSH and executes it on the target, which is expected for remote auditing but means the script will run with the remote user's privileges and have full read access to the target .openclaw directory.
Install Mechanism
No installation step or third-party downloads. The skill is instruction-first with bundled scripts; nothing writes or fetches code from external URLs during install.
Credentials
No environment variables or credentials are required. The script optionally respects OPENCLAW_DIR to target a custom path. There are no requests for unrelated cloud/API keys or system secrets in the provided code.
Persistence & Privilege
always:false and no evidence of the skill modifying other skills or system-wide agent settings. However the tool exposes a --fix mode (passed from the bash wrapper to the Python script) that may perform remediation; README claims fixes require confirmation but the wrapper simply forwards the flag. You should confirm how --fix behaves (dry-run vs destructive) before running it, especially on production instances.
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 clawzembic - 安装完成后，直接呼叫该 Skill 的名称或使用
/clawzembic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 — Lighthouse-style audit for OpenClaw efficiency
元数据
常见问题
Clawzempic 是什么？
Lighthouse-style efficiency audit for OpenClaw. Scores your instance A+ to F across 6 categories (context injection, cron health, session bloat, config, skil... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 581 次。
如何安装 Clawzempic？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install clawzembic」即可一键安装，无需额外配置。
Clawzempic 是免费的吗？
是的，Clawzempic 完全免费（开源免费），可自由下载、安装和使用。
Clawzempic 支持哪些平台？
Clawzempic 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 Clawzempic？
由 anna-claudette（@anna-claudette）开发并维护，当前版本 v1.0.0。
推荐 Skills