Trust Assessment
onedrive received a trust score of 86/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.
SkillShield's automated analysis identified 1 finding: 0 critical, 1 high, 0 medium, and 0 low severity. Key findings include Potential Command Injection and Path Traversal via `curl` arguments.
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 13, 2026 (commit 13146e6a). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings1
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| HIGH | Potential Command Injection and Path Traversal via `curl` arguments The skill demonstrates the use of `curl` commands to interact with OneDrive. These commands involve arguments like filenames (`localfile.txt`, `downloaded.txt`) and item IDs (`{itemId}`) which are likely to be dynamically generated or provided by user input when the skill is used by an AI agent. Specifically, the `--data-binary @localfile.txt` argument for file upload and `-o downloaded.txt` for file download imply reading from and writing to the local filesystem. If these arguments are not properly sanitized or escaped before being interpolated into the shell command, an attacker could inject arbitrary shell commands (e.g., `$(rm -rf /)`) or perform path traversal (e.g., `../../../../etc/passwd`), leading to unauthorized file access, modification, or execution. The `{itemId}` placeholder in the download URL also presents a potential injection point if not sanitized. When constructing shell commands with dynamic arguments, ensure all user-provided or dynamically generated inputs are strictly validated and properly escaped. Consider using a library or framework that safely handles shell command execution (e.g., by passing arguments as a list rather than a single string to `subprocess.run` in Python, or by using `exec.Command` in Go). For file paths, restrict access to a designated sandbox directory and validate paths to prevent traversal (e.g., `path.join` and `os.path.abspath` checks). For URL parameters, ensure proper URL encoding. | LLM | SKILL.md:14 |
Scan History
Embed Code
[](https://skillshield.io/report/78da91f1ae26d567)
Powered by SkillShield