Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Skill Optimizer
v1.0.1［何时使用］当用户需要评估技能质量时；当用户说"检查这个 skill"时；当创建或修改 skill 后需要验证时；当检测到"skill 优化""skill 评估""技能检查"等关键词时；当批量检查多个技能时
⭐ 0· 62·0 current·0 all-time
by@lj22503
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
Name, description, SKILL.md and the included Python scripts all implement a 'skill optimizer' that reads SKILL.md files and scores/validates them. That capability is coherent with the stated purpose. However the package references several specific workspace paths (e.g., ~/.openclaw/workspace/..., /home/admin/.openclaw/...) and includes references to files (references/examples.md, references/scoring.md) that are not present in the bundle — an inconsistency between claimed resources and provided files.
Instruction Scope
SKILL.md and scripts instruct the agent to read and traverse skill directories and open SKILL.md files. The included scripts implement a mandatory template-file check and, when run in batch mode, will walk an entire skills root (hard-coded default path in check-templates.py). This is expected for a validator, but the SKILL.md declares 'templates/references/scripts mentioned must exist' while the bundle itself references references/examples.md and references/scoring.md that are missing — meaning the skill will report failures for its own content. Also the batch default path (/home/admin/.openclaw/.../betterlife/skills and several ~/.openclaw paths) are project-specific and could cause broad scanning of the user's workspace if run without care.
Install Mechanism
No install spec — instruction-only plus scripts. This is low-risk compared to arbitrary remote installs; nothing is downloaded or executed automatically from external URLs.
Credentials
The skill requests no environment variables or credentials. The scripts do rely on specific filesystem paths (e.g., STANDARD_FILE pointing at ~/.openclaw/workspace/docs/SKILL-STANDARD-v3.md) but do not request secrets. Still, hard-coded paths increase the chance of scanning unintended directories.
Persistence & Privilege
The skill does not request elevated privileges, does not set always:true, and does not attempt to modify other skills or system-wide config. It only reads files and prints reports. Running its scripts can exit with non-zero codes but they don't persist configuration or credentials.
What to consider before installing
This skill appears to be a legitimate skill-validator, but it has sloppy packaging and risky defaults you should address before running it broadly: 1) The SKILL.md references reference files (references/examples.md, references/scoring.md) that are not included — the validator will mark the skill as failing for missing files (self-inconsistency). Either add the missing reference files or remove those references. 2) Inspect the two Python scripts (scripts/check-templates.py and scripts/optimize-skill.py) — they contain hard-coded workspace paths (e.g., /home/admin/.openclaw/..., ~/.openclaw/...). Run the scripts with an explicit single-skill argument rather than letting them run in batch mode, or edit the defaults to point to a safe test directory. 3) Run the scripts in a sandbox or with limited permissions first to confirm what directories they read. 4) If you plan to let the agent invoke this skill autonomously, be aware it can enumerate and read SKILL.md and other files in your skills workspace — do not install or enable it if you are unwilling to allow that level of file-read access. If you want to proceed, fix the missing referenced files and adjust/remove hard-coded paths for safer operation.Like a lobster shell, security has layers — review code before you run it.
latestvk97akw5vcg9e6wy5gjgfgxejz983bbwr
License
MIT-0
Free to use, modify, and redistribute. No attribution required.