Trust Assessment
agdr-decide received a trust score of 85/100, 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 Agent instructed to execute arbitrary shell commands.
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 April 1, 2026 (commit 3186bc90). 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 | Agent instructed to execute arbitrary shell commands The skill explicitly provides a bash script within a markdown code block and instructs the agent to execute it to determine the next AgDR file number. This grants the agent the capability to execute arbitrary shell commands. While the provided script itself is relatively robust against injection from filenames due to strict regex parsing (limiting extracted values to digits), the precedent of instructing the agent to execute shell commands creates a significant command injection vulnerability. A malicious user prompt could later instruct the agent to generate and execute a different, harmful shell script, potentially leading to data exfiltration, system modification, or denial of service. 1. Avoid providing direct shell scripts for execution within skill definitions. Instead, define specific, parameterized functions or tools that the agent can call for necessary operations. 2. If shell execution is absolutely unavoidable, ensure it is performed within a highly restricted and sandboxed environment with minimal permissions. 3. Implement strict input validation and sanitization for any variables or user-controlled data passed into shell commands. 4. Consider using built-in agent capabilities or a dedicated, secure API for file system operations (like listing and numbering files) instead of raw shell commands. | LLM | SKILL.md:40 |
Scan History
Embed Code
[](https://skillshield.io/report/3ab9777fa90bc090)
Powered by SkillShield