wechat-miniprogram
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data from external sources, which serves as a significant attack surface.
- Ingestion points: The agent reads project source code, configuration files (e.g.,
project.config.json,app.json), simulator console logs (get_simulator_console), and network traces (get_simulator_network). - Boundary markers: There are no explicit structural boundary markers (such as XML tags or unique delimiters) specified to separate untrusted external data from system instructions.
- Capability inventory: The skill allows for high-impact actions including file system modifications (config editing), cloud database writes (
cloud_db_write_doc), storage management (cloud_manage_storage), and runtime code execution (automation_evaluate). - Sanitization: The instructions recommend manual filtering and de-identification of logs and data before processing, but the agent lacks automated sanitization for the inputs it reads.
- [DYNAMIC_EXECUTION]: The skill implements tools that execute code or expressions dynamically.
- Simulator execution: The
automation_evaluatetool executes arbitrary JavaScript strings (fnSource) within the Mini Program's runtime context. This is intended for debugging but can be abused if the input source is compromised. - Log filtering: The
get_simulator_consoleandget_simulator_networktools accept acommandparameter which is agrepexpression passed to the underlying CLI. Maliciously crafted expressions could potentially lead to command injection if not handled carefully by thewechatideutility. - [EXTERNAL_DOWNLOADS]: The skill references and downloads tools from external sources.
- Trusted sources: It provides links to official documentation at
developers.weixin.qq.comand identifies the official WeChat CI repository atgithub.com/wechat-miniprogram/miniprogram-ci-dist. These are recognized as well-known and trusted services. - [COMMAND_EXECUTION]: The skill interacts with the local environment via the
wechatideCLI andminiprogram-citools to perform project management, compilation, and cloud deployments.
Audit Metadata