wxmini-security-audit
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes source code from WeChat mini-programs provided by the user. If the code being audited contains adversarial instructions (e.g., hidden in comments or strings), it could attempt to manipulate the behavior or output of the analysis agents during the parallel analysis phases.
- Ingestion points: Decompiled JavaScript and JSON files from the mini-program directory are read and passed to the SecretScanner, EndpointMiner, and VulnAnalyzer agents.
- Boundary markers: The agent configurations do not utilize explicit delimiters or instructions to ignore potential commands embedded within the source code data.
- Capability inventory: The skill possesses the capability to read from the local file system, write analysis reports to an output directory, and execute local Python scripts and binaries.
- Sanitization: There is no evidence of sanitization or filtering of the extracted code context before it is submitted to the LLM agents for processing.
- [EXTERNAL_DOWNLOADS]: The framework relies on an external third-party binary,
unveilr.exe, for the initial decompilation phase. - Evidence: The setup instructions guide the user to download the binary from an external GitHub repository (
nicholaschan23/unveilr) and place it within the skill's tools directory for execution.
Audit Metadata