AI Skill Shield scan report: alirezarezvani/claude-skills (846 skills)

Author: ai-skill-shieldCreated Aug 29, 2026Updated Aug 29, 2026

AI Skill Shield

AI Skill Shield scan updated

AI Skill Shield

Field Result
Commit 19392f7a0826
Score 81/100
Status ⚠️ Review recommended
Findings 10 critical · 946 high · 2865 medium · 2173 low
Scanned 2026-08-29

View the full AI Skill Shield report

Findings summary

Critical — curl sending data to external

Location: engineering-team/playwright-pro/skills/report/SKILL.md:91

curl with POST/data flags can exfiltrate file contents or environment data to remote servers.

Recommendation: Block outbound data transfer commands unless explicitly required and documented.

Critical — System prompt extraction

Location: engineering-team/skills/ai-security/SKILL.md:117

Attempts to extract the system prompt. This is a prompt injection technique to understand defenses.

Recommendation: Never reveal system prompt content. Implement prompt injection detection.

Critical — Download then execute pattern

Location: engineering-team/skills/senior-ml-engineer/SKILL.md:59

Downloads a file and executes it. This is a staged malware delivery pattern.

Recommendation: Never download and execute files. Verify and sandbox any downloaded binaries.

Critical — Database drop/delete command

Location: engineering/caveman/skills/caveman/SKILL.md:57

Destructive database command that can cause permanent data loss.

Recommendation: Ensure destructive SQL is only executed in controlled migrations with backups.

Critical — Destructive delete on root

Location: engineering/docker-development/skills/docker-development/SKILL.md:337

Recursive force delete targeting root directory is extremely destructive.

Recommendation: Never delete from root. Use safer file operations with proper path validation.

Critical — SS-SHELL-001

Location: engineering/docker-development/skills/docker-development/SKILL.md:337

Dangerous rm -rf / command that recursively deletes all files

Critical — Direct exec() call

Location: engineering/security-guidance/skills/security-guidance/SKILL.md:22

exec() executes shell commands and is a major RCE vector. Use execFile or spawn with shell:false instead.

Recommendation: Replace exec() with execFile() or use spawn() with shell: false

Critical — Database drop/delete command

Location: engineering/skills/pr-review-expert/SKILL.md:206

Destructive database command that can cause permanent data loss.

Recommendation: Ensure destructive SQL is only executed in controlled migrations with backups.

Critical — SSH private key access

Location: engineering/skills/skill-security-auditor/SKILL.md:121

SSH private key files grant server access. Their exposure is a critical security incident.

Recommendation: Never read SSH private key files. Use SSH agent or keychain instead.

Critical — curl sending data to external

Location: research/grants/skills/grants/SKILL.md:128

curl with POST/data flags can exfiltrate file contents or environment data to remote servers.

Recommendation: Block outbound data transfer commands unless explicitly required and documented.

High — Missing required field: name

Location: .gemini/skills/README/SKILL.md:1

Frontmatter is missing the required "name" field

Recommendation: Add "name" to the frontmatter section

High — Missing required field: description

Location: .gemini/skills/README/SKILL.md:1

Frontmatter is missing the required "description" field

Recommendation: Add "description" to the frontmatter section

Showing the 12 highest-priority findings. View the full report for 8250 more.

Add this status badge to the README

AI Skill Shield

Copy this Markdown into the README:

markdown
[![AI Skill Shield](https://ai-skill-shield.suppeng.com/api/badge/github/alirezarezvani/claude-skills?v=2)](https://ai-skill-shield.suppeng.com/trust/github/alirezarezvani/claude-skills)

This issue is maintained automatically after new default-branch scans. Automated scan results are evidence for review, not a guarantee of safety.

Source: alirezarezvani/claude-skills