Skillv1.0.0

ClawScan security

Luke Find Skills · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

ReviewMar 14, 2026, 7:41 AM

Verdict: Review
Confidence: medium
Model: gpt-5-mini
Summary: The skill's purpose (finding and installing other skills) is coherent, but the instructions encourage installing arbitrary third-party packages via npx (including global, unattended installs) and the package metadata has inconsistencies; these issues warrant caution before installing or following its automated install suggestions.
Guidance: This skill is broadly coherent for finding skills, but exercise caution before following its install recommendations: - The instructions rely on npx/npm but the manifest does not declare that requirement — ensure your environment has npm/npx and understand that npx will download and run remote code. - The SKILL.md suggests 'npx skills add <pkg> -g -y' which installs packages globally and skips confirmation; avoid unattended global installs. Prefer installing locally or without '-y' so you can inspect prompts. - Before adding a skill, review its source (GitHub repo/package page) and check the publisher/owner. The provided _meta.json ownerId and the registry ownerId in the listing do not match and the versions differ — that metadata mismatch reduces trust. - If you plan to install a discovered skill, review its repository and package contents, run installs in a sandbox or VM if possible, and prefer installing without '-g' and without skipping confirmations. If you want a higher-confidence assessment, provide the specific skill package names/URLs you plan to install or the contents of the target skill repositories so their install behavior and code can be reviewed.

Review Dimensions

Purpose & Capability: noteThe skill claims to help discover and install agent skills and its SKILL.md describes using the 'npx skills' CLI to find and add skills — that matches the stated purpose. However, the manifest does not declare that 'npx' / npm is required even though the runtime instructions rely on it, and the SKILL.md encourages global unattended installs (npx skills add -g -y), which expands the scope beyond a simple finder.
Instruction Scope: concernInstructions tell the agent/user to run 'npx skills find' and 'npx skills add <owner/repo@skill>', and explicitly recommend '-g -y' to install globally and skip confirmations. That directs the agent to fetch and execute third-party code from external sources without requiring or recommending prior review, which increases risk. The instructions do not advise verifying package provenance or reviewing code before installing.
Install Mechanism: noteThere is no install spec (instruction-only), which is low-risk in itself. But the SKILL.md's recommended workflow uses 'npx skills add' to download and install skills from remote sources; npx/npm installs can execute arbitrary code from registries or repositories. The SKILL.md points to https://skills.sh/ (a reasonable central index), but does not constrain installs to vetted sources or recommend reviewing packages first.
Credentials: okThe skill requests no environment variables, credentials, or config paths, which is proportionate for a discovery helper. There is no direct request for secrets or unrelated credentials.
Persistence & Privilege: okalways is false and there is no install-time code or persistent privileges requested by the skill itself. The SKILL.md's recommended use of global installs for other skills increases system-wide impact, but that is an implication of the recommended workflow rather than a property of this skill's manifest.