Back to skill
Skillv0.1.1

VirusTotal security

ZugaShield Security Scanner · External malware reputation and Code Insight signals for this exact artifact hash.

Scanner verdict

BenignApr 30, 2026, 4:07 AM
Hash
1024a307c380bc4731e7db47344fe76686be52c6cd496690e77f2ac6c4c48239
Source
palm
Verdict
benign
Code Insight
Type: OpenClaw Skill Name: zugashield Version: 0.1.1 The OpenClaw AgentSkills skill bundle 'zugashield' is a security scanner designed to protect AI agents from various attacks like prompt injection, SSRF, command injection, and data leakage. The code demonstrates strong security practices, such as whitelisting environment variables passed to child processes (`src/preflight.ts`, `src/shield-client.ts`) to prevent secret leakage, and implementing fail-closed mechanisms, especially for tool execution (`src/hooks/pre-tool-exec.ts`). The `SKILL.md` and `README.md` are purely descriptive and do not contain prompt injection attempts against the agent. Crucially, the `test/integration.test.ts` file includes extensive tests against a wide range of real attack payloads, with the explicit expectation that these attacks *will be blocked*, strongly indicating a security-focused and non-malicious intent. There is no evidence of intentional harmful behavior, data exfiltration, backdoors, or obfuscation.
External report
View on VirusTotal