Security Audit
PabloLION/bmad-plugin:plugins/bmad/skills/bmad-agent-builder
github.com/PabloLION/bmad-pluginTrust Assessment
PabloLION/bmad-plugin:plugins/bmad/skills/bmad-agent-builder 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 12 findings: 4 critical, 7 high, 0 medium, and 1 low severity. Key findings include Arbitrary command execution, Unsafe deserialization / dynamic eval, Dangerous call: subprocess.run().
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The Manifest Analysis layer scored lowest at 0/100, indicating areas for improvement.
Last analyzed on April 11, 2026 (commit 17efb6ce). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings12
| 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 | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:525 | |
| 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 | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:527 | |
| 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 | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:529 | |
| 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 | plugins/bmad/skills/bmad-agent-builder/scripts/scan-scripts.py:38 | |
| HIGH | Unsafe deserialization / dynamic eval Decryption followed by code execution Remove obfuscated code execution patterns. Legitimate code does not need base64-encoded payloads executed via eval, encrypted-then-executed blobs, or dynamic attribute resolution to call system functions. | Manifest | plugins/bmad/skills/bmad-agent-builder/scripts/scan-scripts.py:8 | |
| HIGH | Unsafe deserialization / dynamic eval Decryption followed by code execution Remove obfuscated code execution patterns. Legitimate code does not need base64-encoded payloads executed via eval, encrypted-then-executed blobs, or dynamic attribute resolution to call system functions. | Manifest | plugins/bmad/skills/bmad-agent-builder/scripts/scan-scripts.py:367 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function 'main'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:525 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function 'main'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:527 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function 'main'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | plugins/bmad/skills/bmad-agent-builder/scripts/generate-html-report.py:529 | |
| HIGH | Dangerous call: subprocess.run() Call to 'subprocess.run()' detected in function '_run_command'. This can execute arbitrary code. Avoid using dangerous functions like exec/eval/os.system. Use safer alternatives. | Static | plugins/bmad/skills/bmad-agent-builder/scripts/scan-scripts.py:38 | |
| HIGH | Unpinned npm package execution via npx The script uses `npx` to execute `@biomejs/biome` without specifying a version. This means `npx` will always fetch and run the latest available version of the package. This is a supply chain risk as a malicious update to `@biomejs/biome` could be automatically downloaded and executed, compromising the system. Pin the version of `@biomejs/biome` when using `npx`. For example, `npx --yes @biomejs/biome@1.5.3 check ...` to ensure a specific, vetted version is used. Regularly review and update pinned versions. | Static | scripts/scan-scripts.py:307 | |
| LOW | Potential for arbitrary file read by analysis scripts Several Python scripts (`generate-html-report.py`, `prepass-execution-deps.py`, `prepass-prompt-metrics.py`, `prepass-structure-capabilities.py`, `scan-path-standards.py`, `scan-scripts.py`) are designed to read file content from paths provided as command-line arguments. If the agent's logic allows untrusted user input to directly influence these file paths, it could lead to the reading of arbitrary files on the system, potentially exfiltrating sensitive data. While these scripts are intended for analyzing skill files, the host LLM should ensure proper sandboxing and validation of input paths. The host LLM environment should implement strict input validation and sandboxing for any file paths passed to these scripts, ensuring they only operate within the designated skill directory or other safe, restricted areas. Avoid passing user-controlled strings directly as file paths without sanitization. | Static | scripts/generate-html-report.py:21 |
Scan History
Embed Code
[](https://skillshield.io/report/269d94715cc82493)
Powered by SkillShield