Trust Assessment
skill-creator received a trust score of 10/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.
SkillShield's automated analysis identified 8 findings: 3 critical, 4 high, 0 medium, and 1 low severity. Key findings include File read + network send exfiltration, Sensitive path access: AI agent config, Potential Command Injection via `evaluate_skill` tool parameter.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The Manifest Analysis layer scored lowest at 10/100, indicating areas for improvement.
Last analyzed on April 1, 2026 (commit d7538ceb). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings8
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | File read + network send exfiltration AI agent config/credential file access Remove access to sensitive files not required by the skill's stated purpose. SSH keys, cloud credentials, and browser data should never be read by skills unless explicitly part of their declared functionality. | Manifest | skills/skill-creator/SKILL.md:132 | |
| CRITICAL | File read + network send exfiltration AI agent config/credential file access Remove access to sensitive files not required by the skill's stated purpose. SSH keys, cloud credentials, and browser data should never be read by skills unless explicitly part of their declared functionality. | Manifest | skills/skill-creator/SKILL.md:152 | |
| CRITICAL | File read + network send exfiltration AI agent config/credential file access Remove access to sensitive files not required by the skill's stated purpose. SSH keys, cloud credentials, and browser data should never be read by skills unless explicitly part of their declared functionality. | Manifest | skills/skill-creator/SKILL.md:156 | |
| HIGH | Sensitive path access: AI agent config Access to AI agent config path detected: '~/.claude/'. This may indicate credential theft. Verify that access to this sensitive path is justified and declared. | Static | skills/skill-creator/SKILL.md:132 | |
| HIGH | Sensitive path access: AI agent config Access to AI agent config path detected: '~/.claude/'. This may indicate credential theft. Verify that access to this sensitive path is justified and declared. | Static | skills/skill-creator/SKILL.md:152 | |
| HIGH | Sensitive path access: AI agent config Access to AI agent config path detected: '~/.claude/'. This may indicate credential theft. Verify that access to this sensitive path is justified and declared. | Static | skills/skill-creator/SKILL.md:156 | |
| HIGH | Potential Command Injection via `evaluate_skill` tool parameter The `SKILL.md` describes an `evaluate_skill` tool that runs a Python script (`run_loop.py`) and accepts parameters, including `eval_set_path`. If the agent framework allows user-provided input to populate `eval_set_path` without proper sanitization, an attacker could inject arbitrary shell commands (e.g., `eval_set_path="malicious.json; rm -rf /"`). This vulnerability exists in the interaction between the agent framework and the described tool, where the skill defines the interface that could be exploited if the tool invocation mechanism is not secure. The agent framework responsible for invoking `evaluate_skill` must sanitize or validate all user-provided inputs passed to external command-line tools. Specifically, ensure `eval_set_path` is validated to be a safe file path and does not contain shell metacharacters before being used in a subprocess call. Consider using a dedicated API for tool invocation instead of direct shell execution where possible. | LLM | SKILL.md:199 | |
| LOW | Unpinned dependency for `anthropic` package The `evaluate_skill` tool description mentions a requirement for the `anthropic` Python package but does not specify a version. This unpinned dependency can lead to unexpected behavior, compatibility issues, or security vulnerabilities if a future version of the package introduces breaking changes or security flaws. While `anthropic` is a first-party library, relying on the latest version without pinning can introduce instability. Specify a precise version or a version range for the `anthropic` package in the `evaluate_skill` tool's dependency management (e.g., `anthropic==0.1.2` or `anthropic>=0.1.2,<0.2.0`). This ensures consistent behavior and reduces the risk of unexpected issues from future package updates. | LLM | SKILL.md:194 |
Scan History
Embed Code
[](https://skillshield.io/report/9e82664efade5d23)
Powered by SkillShield