java-audit-skill
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands (find, grep, wc, xargs) and specialized security tools (ripgrep, Semgrep) through its scripts and instructions. These operations are restricted to the local workspace for the purpose of static code analysis and reporting.
- [EXTERNAL_DOWNLOADS]: The framework suggests using web_search to query official security databases like NVD, CVE, and GitHub Advisories to verify vulnerability data for identified dependencies. It also references official repositories for security guidelines (Oracle, OWASP, Tencent). All referenced sources are well-known security entities and fall within trusted scopes for research.
- [INDIRECT_PROMPT_INJECTION]: As a tool that ingests untrusted Java/Kotlin source code, the skill is theoretically vulnerable to indirect prompt injection (e.g., malicious code comments designed to mislead the auditor). However, the skill explicitly mandates '5 Anti-Hallucination Iron Rules' that require the agent to verify all findings against actual file reads and line numbers, effectively mitigating this risk.
- [CREDENTIALS_UNSAFE]: The skill provides rules and patterns specifically designed to find hardcoded credentials and insecure configurations in target codebases. The skill itself does not contain any hardcoded secrets or insecure credential management practices.
- [DATA_EXFILTRATION]: While the skill performs comprehensive local scanning, it lacks any mechanisms to exfiltrate data to unauthorized remote servers. Audit results are written to a local 'audit-output' directory within the project path.
Audit Metadata