📦 Identity Guard — 实用工具
v0.0.1安全性 CHECKPOINT. Trigger 技能 IMMEDIATELY whenever 用户 asks about owner/master, their personal 信息, plans, activities, or...
0· 275·0 当前·0 累计
by @zhongxuyang 安全扫描
OpenClaw
安全
high confidenceThe skill's files and runtime instructions are consistent with its stated purpose (verifying users by sender_id); there are no network endpoints, secrets requests, or surprising installs, though a couple of small operational caveats exist.
评估建议
This skill appears to do exactly what it says: block sensitive owner/master queries unless the sender_id is allowlisted. Before installing, confirm: (1) whether you trust the agent to write/modify identities.json via chat (init flow) — if not, use the CLI init.sh instead; (2) your runtime environment has ripgrep (rg) and python3 if you plan to use whoami.sh or add-user.sh; (3) you are comfortable with whoami.sh reading local OpenClaw session logs (~/.openclaw/.../sessions) to discover sender_id....详细分析 ▾
✓ 用途与能力
Name/description (identity verification by sender_id) matches the included scripts and SKILL.md. The scripts implement an allowlist/master_id model (guard.sh, init.sh, add-user.sh) and helper whoami.sh to extract sender_id from local session logs. No unrelated credentials, network calls, or external services are requested.
ℹ 指令范围
SKILL.md's runtime instructions stay focused on verifying sender_id before answering sensitive queries. It explicitly forbids memory-based verification and data disclosure. One notable instruction: the assistant is allowed to initialize identities.json via chat (in DMs only) — that implies the assistant may be instructed to edit a local config file. This is intentional for the skill but means the agent must have write access to the skill workspace for that flow to work.
✓ 安装机制
No install script or external downloads are used; the skill is instruction-only with included shell/python scripts. Nothing is fetched from arbitrary URLs and no archives are extracted.
ℹ 凭证需求
The skill declares no required environment variables or credentials, which aligns with its purpose. Minor mismatch: scripts assume utilities that were not declared as required — whoami.sh requires ripgrep (rg) and add-user.sh requires python3. whoami.sh reads session logs under ~/.openclaw which is necessary for its function but does access local session files (sender_id values).
✓ 持久化与权限
Skill is not forced-always; it is user-invocable and can be invoked autonomously (platform default). It does not request elevated system-wide privileges and only reads/writes its own identities.json in the skill workspace when run. The chat-based init feature requires the agent to be permitted to modify that file; confirm whether you want the agent to have that capability.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
安装命令
点击复制官方npx clawhub@latest install identity-guard
镜像加速npx clawhub@latest install identity-guard --registry https://cn.longxiaskill.com