70High
The skill is vulnerable to command injection and relies on external
Behavior Analysis
Description verified — behavior matches claim
Claims to do
Check: Review Before You Ship: Prefix your first line with 🥷 inline, not as its own paragraph.
Actually does
This skill uses `git` and `gh` CLI to review code diffs, triage GitHub issues and PRs, and manage releases. It reads repository context (e.g., READMEs, manifests, CI configs), performs checks for hard stops and scope drift, and can apply safe auto-fixes. It interacts with GitHub to list/view/close issues and PRs, push commits, create releases, and add reactions, and can execute project-specific verification commands.
Skill Info
Nametw93/waza/check
Registrygithub
Versionb89a98a9e69e
PURLpkg:github/tw93/waza@b89a98a9e69e?skill=check
Stars4,497
Installs3,900
Source
SHA-25677b4638088cf...
SHA-17fae4eea0db1...
MD53f9edf5cf669...
TLSH3842e65b7e99...
ssdeep192:omCX7N4r...
Size12,986 bytes
Install
Assessments (3)
AI Agent Traps ↗Command Injection via User-Controlled Inputhighcommand execution
The skill constructs shell commands using potentially user-controlled input (e.g., issue keywords, search terms) without explicit sanitization, creating a command injection vulnerability. An attacker could craft malicious input to execute arbitrary commands.
100% confidenceline 55
`git log --oneline <latest-tag>..HEAD | grep -i "<keyword>"` `grep -r "name" .`
Broad Command Execution Surfacemediumcommand execution
The skill has extensive capabilities to execute various external commands (`gh`, `git`, `bash scripts/run-tests.sh`). This broad access, while functional, presents a significant attack surface if the agent's reasoning or input is compromised, enabling a wide range of malicious actions.
90% confidenceline 128
`gh issue list`, `gh pr list`, `git log`, `gh api`, `bash scripts/run-tests.sh`, `grep -r`, `vercel env ls`, `git remote -v`
Extensive Reconnaissance and PII Detection Capabilitiesmediumreconnaissance
The skill performs wide-ranging reconnaissance, including listing issues/PRs, reading various project files, and explicitly detecting PII during document review. If the agent's output or communication channels are compromised, this gathered sensitive information could be exfiltrated.
90% confidenceline 150
`gh issue list`, `gh pr list`, `git log`, `grep -r`, `Privacy scan: Detect PII (names, companies, employment dates, salary hints, location details).`
External Configuration Poisoning & Sub-Agent Spawninghighrag poisoning
The agent's critical decision-making, project context extraction, and sub-agent configurations rely on external files (`references/project-context.md`, `references/persona-catalog.md`). If these files are compromised, an attacker could poison the agent's knowledge base or inject malicious sub-agent prompts.
100% confidenceline 30
Load `references/project-context.md` Load `references/persona-catalog.md` to determine which specialists activate. Launch all activated specialists in parallel...
Unconfirmed Automated Actions (safe_auto)mediumautonomy abuse
The skill is designed to apply `safe_auto` fixes immediately without user confirmation. While intended for 'unambiguous, risk-free' changes, a flaw in this classification or a manipulated definition could lead to unintended or harmful modifications without human oversight.
90% confidenceline 111
| `safe_auto` | Unambiguous, risk-free: typos, missing imports, style inconsistencies | Apply immediately |
Badge
Markdown
[](https://mondoo.com/ai-agent-security/skills/github/tw93/waza/check)HTML
<a href="https://mondoo.com/ai-agent-security/skills/github/tw93/waza/check"><img src="https://mondoo.com/ai-agent-security/api/badge/github/tw93/waza/check.svg" alt="Mondoo Skill Check" /></a>Image URL
https://mondoo.com/ai-agent-security/api/badge/github/tw93/waza/check.svgSecure your AI agents
Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.
- Continuous scanning of skills across all registries
- Policy enforcement before skills reach your agents
- Integration with your existing security workflow