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 7 findings: 4 critical, 3 high, 0 medium, and 0 low severity. Key findings include Arbitrary command execution, Dangerous call: subprocess.run(), Dangerous call: subprocess.Popen().
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 July 1, 2026 (commit 35414756). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings7
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Arbitrary command execution Python shell execution (os.system, subprocess) Review all shell execution calls. Ensure commands are static (not built from user input), use absolute paths, and are strictly necessary. Prefer library APIs over shell commands. | Manifest | skills/skill-creator/eval-viewer/generate_review.py:291 | |
| CRITICAL | Arbitrary command execution Python shell execution (os.system, subprocess) Review all shell execution calls. Ensure commands are static (not built from user input), use absolute paths, and are strictly necessary. Prefer library APIs over shell commands. | Manifest | skills/skill-creator/scripts/improve_description.py:35 | |
| CRITICAL | Arbitrary command execution Python shell execution (os.system, subprocess) Review all shell execution calls. Ensure commands are static (not built from user input), use absolute paths, and are strictly necessary. Prefer library APIs over shell commands. | Manifest | skills/skill-creator/scripts/run_eval.py:85 | |
| CRITICAL | Indirect Prompt Injection leading to Remote Code Execution (RCE) via claude -p The `improve_description.py` and `run_eval.py` scripts execute `claude -p` (Claude Code) with untrusted inputs directly embedded in the prompt or query. Specifically, `improve_description.py` embeds the raw content of `SKILL.md` into the prompt, and `run_eval.py` passes untrusted queries from `evals.json` directly to `claude -p`. Because Claude Code is an agentic LLM with access to powerful system tools (such as executing bash commands and reading/writing files), a malicious skill can exploit this via indirect prompt injection to execute arbitrary shell commands on the user's host system. Do not run agentic LLM sessions (like `claude -p` with tool access) on untrusted inputs such as third-party `SKILL.md` files or `evals.json` queries. If analysis or evaluation is required, use a non-agentic LLM API call without system tool access, or run the agent in a secure, isolated sandbox container. | LLM | scripts/improve_description.py:104 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function '_kill_port'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | skills/skill-creator/eval-viewer/generate_review.py:291 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function '_call_claude'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | skills/skill-creator/scripts/improve_description.py:35 | |
| HIGH | Dangerous call: subprocess.Popen() Call to 'subprocess.Popen()' detected in function 'run_single_query'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | skills/skill-creator/scripts/run_eval.py:85 |
Scan History
Embed Code
[](https://skillshield.io/report/1715b8e3f2f39fcf)
Powered by SkillShield