skillguard-check

PassAudited by ClawScan on May 10, 2026.

Overview

This is a coherent security-check skill, but users should know it runs a local script and sends installed skill names to skillguard.vip for lookup.

This skill appears reasonable for checking installed skills against a public security database. Before installing, decide whether you are comfortable with a local script enumerating your skill directory names and sending those slugs to skillguard.vip; use a pinned or verified install source if installing from GitHub.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Low

#ASI02: Tool Misuse and Exploitation

What this means

The agent may run a local Python checker when a skill audit is relevant; the reviewed code does not modify files but does perform directory discovery and web lookups.

Why it was flagged

The skill allows Bash and instructs the agent to run its bundled Python script. This is expected for a local audit helper and the script source is provided, but it is still local command execution.

Skill content

allowed-tools:
  - Bash
  - Read
...
python3 scripts/check.py

Recommendation

Review the bundled script before installation if you are cautious about local command execution, and invoke it only when you want a skill audit.

Low

#ASI07: Insecure Inter-Agent Communication

What this means

skillguard.vip can learn the names of installed skills being checked, including private or unpublished skill directory names if present.

Why it was flagged

The skill discloses that local skill names are transmitted to an external provider. This is purpose-aligned and limited to slugs, but it is still a privacy-relevant data flow.

Skill content

Each name (slug) is sent as a path component to skillguard.vip in a plain HTTPS GET. No headers identify the user; standard Cloudflare access logs apply.

Recommendation

Use this only if you are comfortable sharing installed skill slugs with skillguard.vip, or use the documented self-hosted API option if you need stricter privacy.

Info

#ASI04: Agentic Supply Chain Vulnerabilities

What this means

If installed using those README commands later, the downloaded files could differ from the reviewed artifact version.

Why it was flagged

The README provides user-directed installation commands that fetch from GitHub's current main branch without pinning a commit or checksum. This is not automatic in the provided install spec, but it is a supply-chain consideration.

Skill content

git clone --depth 1 --filter=blob:none --sparse https://github.com/yangyixxxx/skillguard.git skillguard-check-tmp
...
curl -fsSLo ~/.claude/skills/skillguard-check/SKILL.md https://raw.githubusercontent.com/yangyixxxx/skillguard/main/skills/skillguard-check/SKILL.md

Recommendation

Prefer installing the reviewed marketplace artifact, or pin a specific commit and verify file contents when using the GitHub/curl installation path.