MondooMondoo
AI Agent Security
Skills
Log inGet Assessment

AI Agent Skill Check is a free AI agent skill security scanner by Mondoo. We scan skills across ClawHub, Skills.sh, GitHub, Claude Marketplace, and SkillsMP to detect prompt injection, credential theft, data exfiltration, agent impersonation, and 28 threat types before they reach your agents.

Mondoo

  • Vulnerability Management
  • Technology
  • Services

Solutions

  • Financial Services
  • Manufacturing
  • Healthcare

Resources

  • Blog
  • Skill Check
  • Documentation
  • GitHub

Company

  • About
  • Careers
  • Partners
  • Contact

Legal

  • Privacy
  • Terms
  • Imprint
MondooMondoo© 2026 Mondoo, Inc.

Skills

Browse, search, and filter AI agent skills across all registries.

Description Mismatch
SkillAI AgentsSummaryStarsDownloadsFindingsRisk
agent-skills/firebase-security-rules-auditor
firebase
GitHubClaude CodeGemini CLICursor

The skill misrepresents itself as an active Firebase security rules

224455.6k1
5Low
agent-skills/firebase-app-hosting-basics
firebase
GitHubClaude CodeGemini CLICursor

The skill deceptively claims to deploy applications but only

224455.6k1
40Medium
claude-code/plugin-structure
anthropics
GitHubClaude Code

This skill describes a plugin architecture

114.4k455.6k4
40Medium
claude-code/hook-development
anthropics
GitHubClaude Code

The skill misrepresents its capabilities,

114.4k455.6k1
5Low
Claude-plugins/tdd
Fleron
GitHubClaude Code

The skill uses dogmatic language to rigidly enforce

0455.6k2
15Low
Claude-plugins/handoff-md
Fleron
GitHubClaude Code

Designed to write a shared document, this skill poses a

0455.6k2
70High
azure-skills/azure-ai
microsoft
GitHubClaude CodeGemini CLI

The skill misrepresents its capabilities,

633174.7k1
70High
azure-skills/azure-storage
microsoft
GitHubClaude CodeGemini CLI

The skill exposes direct Azure CLI

633174.5k2
70High
github-copilot-for-azure/azure-ai
microsoft
GitHubSkills.sh

The skill misrepresents its Azure service support,

186103.8k1
40Medium
github-copilot-for-azure/azure-storage
microsoft
GitHub

The skill's `az storage`

186103.3k2
70High
github-copilot-for-azure/azure-validate
microsoft
GitHub

This skill enables arbitrary command execution and prompt injection, overriding

186103.1k12
100Critical
impeccable/polish
pbakaus
GitHubClaude CodeSkills.sh

The skill deceptively claims to perform

20.1k71.4k2
70High
impeccable/adapt
pbakaus
GitHubClaude Code

The skill's description is misleading, and it introduces

19.5k63.0k2
40Medium
impeccable/colorize
pbakaus
GitHubClaude Code

The skill misrepresents its capabilities, claiming to colorize

19.5k62.3k1
40Medium
impeccable/optimize
pbakaus
GitHubClaude Code

The skill misrepresents its capabilities, claiming to diagnose and

19.5k62.2k1
70High
impeccable/delight
pbakaus
GitHubClaude Code

The skill introduces a supply chain risk by depending on an

19.5k61.5k2
40Medium
word-docx
ivangdavila
OpenClaw

The skill misrepresents its capabilities, claiming to create

26760.5k1
70High
cli/lark-minutes
larksuite
GitHub

The skill risks RAG poisoning via an external `SK

7.9k57.0k3
70High
cli/lark-contact
larksuite
GitHub

The skill is vulnerable to RAG/memory poisoning

7.9k56.8k2
70High
excel-xlsx
ivangdavila
OpenClaw

The skill misrepresents its capabilities, claiming to manipulate

21154.5k1
70High
impeccable/harden
pbakaus
GitHubClaude CodeSkills.sh

The skill misrepresents its hardening capabilities, only offering guidance

20.1k53.0k1
40Medium
powerpoint-pptx
ivangdavila
OpenClaw

The skill falsely advertises PowerPoint editing

10234.1k1
70High
data-analysis
ivangdavila
OpenClaw

The skill misrepresents its capabilities, claiming to

8726.2k1
70High
moltguard
thomaslwang
OpenClaw

The skill deceptively claims security protection, exposes

10722.7k4
40Medium
agent-skills/convex-migration-helper
get-convex
GitHub

The skill deceptively claims to "plan

2221.5k1
40Medium
market-research
ivangdavila
OpenClaw

The skill recommends installing and updating unverified software via

7217.1k3
40Medium
baoyu-skills/baoyu-xhs-images
jimliu
GitHubClaude Code

The skill is vulnerable to path traversal, command

14.6k16.9k6
70High
china-stock-analysis
paulshe
OpenClaw

The skill provides financial advice and uses web search,

4016.8k4
70High
baoyu-skills/baoyu-cover-image
jimliu
GitHubClaude Code

The skill is vulnerable to path traversal and

14.6k16.7k7
40Medium
baoyu-skills/baoyu-infographic
jimliu
GitHubClaude Code

The skill allows arbitrary shell command execution and local file content

14.6k15.9k5
100Critical
baoyu-skills/baoyu-post-to-x
jimliu
GitHubClaude Code

This skill uses system-level tools and broad permissions to

14.6k15.3k9
100Critical
baoyu-skills/baoyu-comic
jimliu
GitHubClaude Code

This skill is vulnerable to arbitrary

14.6k15.3k6
100Critical
image
ivangdavila
OpenClaw

The skill introduces supply chain risks by loading

2514.6k4
40Medium
post-job
zhangdong
OpenClaw

This skill executes arbitrary code, injects prompts into

814.3k16
100Critical
skills/tdd
mattpocock
GitHub

The skill appears safe, though its description misrepresents its capabilities as purely informational rather than active.

15.3k13.8k1
15Low
skills/vitest
antfu
GitHub

The skill deceptively claims active testing utility but only

4.6k13.8k1
40Medium
baoyu-skills/baoyu-format-markdown
jimliu
GitHubClaude Code

The skill deceptively modifies user files and executes

14.6k13.6k3
100Critical
awesome-copilot/refactor
github
GitHub

The skill deceptively claims to refactor code but only

29.9k13.6k1
70High
agents/fastapi-templates
wshobson
GitHub

The skill promotes insecure CORS configuration

33.7k13.5k3
40Medium
obsidian-skills/json-canvas
kepano
GitHubClaude Code

The skill defines attributes allowing arbitrary local file paths, posing

24.3k13.1k2
70High
skills/write-a-prd
mattpocock
GitHub

The skill describes dangerous repo exploration and GitHub

15.3k13.0k4
70High
screenshot
ivangdavila
OpenClaw

The skill misrepresents its functionality, providing instructions on

2912.7k1
70High
zoho-mail
byungkyu
OpenClaw

The skill risks arbitrary code execution via unsanitized

812.6k2
70High
agents/code-review-excellence
wshobson
GitHub

The skill misrepresents itself as

33.7k12.3k1
40Medium
git
ivangdavila
OpenClaw

The skill misrepresents itself as an active Git agent,

2611.7k1
40Medium
eno
wscats
OpenClaw

The skill misrepresents its capabilities, claiming programmatic analysis and

411.0k1
15Low
awesome-copilot/multi-stage-dockerfile
github
GitHub

The skill appears safe, though its description overstates its active role in Dockerfile creation.

29.9k10.9k1
40Medium
docker
ivangdavila
OpenClaw

The skill introduces a supply chain risk by suggesting

2110.7k2
40Medium
Page 1 of 15