Trust Assessment
google-veo received a trust score of 21/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.
SkillShield's automated analysis identified 4 findings: 2 critical, 1 high, 1 medium, and 0 low severity. Key findings include Arbitrary command execution, Remote code execution: curl/wget pipe to shell, Potential Command Injection via User-Controlled Prompt.
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 Findings4
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| CRITICAL | Arbitrary command execution Remote code download piped to interpreter 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 | skills/okaris/google-veo/SKILL.md:9 | |
| CRITICAL | Remote code execution: curl/wget pipe to shell Detected a pattern that downloads and immediately executes remote code. This is a primary malware delivery vector. Never pipe curl/wget output directly to a shell interpreter. | Static | skills/okaris/google-veo/SKILL.md:9 | |
| HIGH | Potential Command Injection via User-Controlled Prompt The skill is declared with broad Bash permissions (`Bash(infsh *)`) and provides examples where user-controlled input (the 'prompt' within the `--input` JSON argument) is directly interpolated into a shell command. If the host LLM constructs this command by directly inserting unescaped user input into the JSON string, an attacker could craft a malicious prompt to break out of the JSON and inject arbitrary shell commands. For example, a prompt like `foo"}' && rm -rf / --input '{ "prompt": "bar` could lead to arbitrary command execution. Implement robust input sanitization and escaping for all user-provided data before it is incorporated into shell commands. Specifically, ensure that the 'prompt' value is properly JSON-escaped and shell-escaped if it's passed as part of a larger shell command. Consider using a more restrictive permission for `Bash` if possible, e.g., `Bash(infsh app run google/veo-* --input *)` to limit the scope of executable commands. | LLM | SKILL.md:13 | |
| MEDIUM | Excessive Bash Permissions Declared The skill declares `Bash(infsh *)` as an allowed tool. This grants permission to execute any command starting with `infsh`. While the examples in `SKILL.md` primarily show `infsh app run`, `infsh app list`, and `infsh app sample`, the broad `*` wildcard allows for potentially destructive or unauthorized `infsh` commands (e.g., `infsh config set`, `infsh app delete-all`) if an attacker can inject arbitrary commands or manipulate the skill's execution flow. This permission is broader than necessary for the skill's stated purpose of generating videos. Narrow the `allowed-tools` permission to the minimum necessary commands. For example, if the skill only needs to run Veo apps, consider a more specific permission like `Bash(infsh app run google/veo-* --input *)` or a list of specific `infsh` subcommands and arguments required for functionality. | LLM | SKILL.md |
Scan History
Embed Code
[](https://skillshield.io/report/72a1513b4652aa6f)
Powered by SkillShield