Trust Assessment
mobile-design received a trust score of 76/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.
SkillShield's automated analysis identified 2 findings: 0 critical, 2 high, 0 medium, and 0 low severity. Key findings include Unjustified 'Bash' permission declared, Potential data exfiltration via sensitive file reading and reporting.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.
Last analyzed on February 20, 2026 (commit e36d6fd3). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| HIGH | Unjustified 'Bash' permission declared The skill declares 'Bash' permission, allowing it to execute arbitrary shell commands. The provided 'SKILL.md' and 'scripts/mobile_audit.py' describe static analysis and design principles, which do not inherently require shell execution. The 'mobile_audit.py' script is a Python script and does not directly utilize Bash. This broad permission introduces a significant risk of command injection if the skill's internal logic were to construct and execute commands based on untrusted user input or file content. Remove the 'Bash' permission unless a specific, justified use case is identified and implemented securely. If shell commands are truly needed, consider using more constrained tools or carefully sanitizing all inputs. | LLM | SKILL.md | |
| HIGH | Potential data exfiltration via sensitive file reading and reporting The 'scripts/mobile_audit.py' script, which is part of the untrusted skill package, is designed to read and analyze code files. With the declared 'Read' permission, the skill can be instructed to read any file accessible to the agent. The skill's documentation ('SKILL.md') explicitly mentions 'Security Sins' like 'Hardcoded secrets' as something to avoid. If the script identifies such secrets in a sensitive file (e.g., configuration files, environment variables) and reports them in its output, this constitutes data exfiltration. While the full reporting logic is truncated, the capability and intent to identify and report security issues are present. Implement strict input validation and sanitization for file paths provided to the skill. Ensure the skill's output is carefully filtered to prevent the accidental disclosure of sensitive data found during analysis. Consider restricting the 'Read' permission scope to specific directories or file types if possible. | LLM | scripts/mobile_audit.py:30 |
Scan History
Embed Code
[](https://skillshield.io/report/0bc497c08efcfb53)
Powered by SkillShield