Security Audit
cursor
github.com/Mrc220/agent_flywheel_clawdbot_skills_and_integrationsTrust Assessment
cursor received a trust score of 25/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.
SkillShield's automated analysis identified 5 findings: 1 critical, 3 high, 1 medium, and 0 low severity. Key findings include File read + network send exfiltration, Sensitive path access: SSH key/config, Skill enables access to sensitive user files.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The LLM Behavioral Safety layer scored lowest at 63/100, indicating areas for improvement.
Last analyzed on June 1, 2026 (commit 6a655802). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings5
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | File read + network send exfiltration SSH key/config 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/cursor/SKILL.md:181 | |
| HIGH | Sensitive path access: SSH key/config Access to SSH key/config path detected: '~/.ssh/config'. This may indicate credential theft. Verify that access to this sensitive path is justified and declared. | Static | skills/cursor/SKILL.md:181 | |
| HIGH | Skill enables access to sensitive user files The `cursor` skill allows opening arbitrary files and folders on the filesystem, including potentially sensitive user configuration files such as `~/.ssh/config`, `~/Library/Application Support/Cursor/User/settings.json`, and `~/Library/Application Support/Cursor/User/keybindings.json`. An attacker could prompt the agent to open these files, allowing the LLM to read and potentially exfiltrate their contents. Implement strict access controls or sandboxing for the `cursor` command, or restrict the paths the agent can open. Warn users about the potential for sensitive file access. | LLM | SKILL.md:20 | |
| HIGH | Potential for command injection and supply chain risk via extension installation The `cursor --install-extension <extension-id>` command allows the agent to install extensions. If an attacker can control the `<extension-id>`, they could potentially install a malicious extension that executes arbitrary code, leading to command injection or other system compromise. This also represents a supply chain risk if the extension source is untrusted or the ID is not properly sanitized. Restrict the agent's ability to install extensions, or whitelist allowed extension IDs. Ensure that the `cursor` CLI properly sanitizes inputs for extension IDs to prevent shell injection and validate extension sources. | LLM | SKILL.md:80 | |
| MEDIUM | Broad command execution capabilities The skill grants the agent the ability to execute the `cursor` CLI with a wide range of arguments, including opening files, folders, and managing extensions. This broad command execution capability means that if the agent is compromised or given malicious instructions, it could perform actions on the system that the user has permissions for, potentially leading to unintended modifications or data exposure. Implement strict input validation and sanitization for all arguments passed to the `cursor` command. Consider sandboxing the execution environment for the agent to limit its impact on the host system. | LLM | SKILL.md:12 |
Scan History
Embed Code
[](https://skillshield.io/report/46165d4bcf16958e)
Powered by SkillShield