Trust Assessment
my-tesla received a trust score of 65/100, placing it in the Caution category. This skill has some security considerations that users should review before deployment.
SkillShield's automated analysis identified 2 findings: 1 critical, 1 high, 0 medium, and 0 low severity. Key findings include Skill can exfiltrate precise vehicle location and raw data, Skill grants extensive control over a physical vehicle.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. The LLM Behavioral Safety layer scored lowest at 55/100, indicating areas for improvement.
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 Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Skill grants extensive control over a physical vehicle The skill provides commands for a wide range of critical vehicle operations, including locking/unlocking doors, adjusting climate, managing charging, opening trunks/windows, activating Sentry Mode, honking the horn, and flashing lights. While many of these actions are gated by a `--yes` confirmation flag, an LLM could be prompted to include this flag, leading to unintended or malicious control over the physical vehicle. This level of control poses significant safety and security risks if misused, as it directly impacts a physical asset. Implement robust, multi-factor human confirmation for all critical vehicle control actions, especially those with safety implications, that cannot be bypassed by an LLM. Consider rate-limiting or geo-fencing for certain commands. Ensure the LLM's access to such powerful tools is strictly controlled and monitored. | LLM | SKILL.md:69 | |
| HIGH | Skill can exfiltrate precise vehicle location and raw data The skill provides commands (`status --json`, `report --json --raw-json`, `location --yes`) that explicitly output raw vehicle data, including precise location and drive state. If an LLM is prompted to execute these commands and then process or summarize the output, this sensitive information could be inadvertently exfiltrated to the LLM's context or external systems. The documentation itself highlights this capability, noting that `status --json` outputs `raw vehicle_data, which may include location/drive_state` and `location --yes` provides precise coordinates. Implement stricter output filtering or redaction for sensitive data by default, especially for location and raw vehicle data, before presenting it to the LLM. Require explicit, high-friction user confirmation for commands that expose precise location or raw vehicle data. Ensure the LLM is explicitly instructed *not* to log or transmit such data. | LLM | SKILL.md:59 |
Scan History
Embed Code
[](https://skillshield.io/report/8404f1d281425034)
Powered by SkillShield