The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using grep to search through the project's source code directory (ruoyi-modules/) based on user-provided function names.
[CREDENTIALS_UNSAFE]: The skill explicitly reads ruoyi-admin/src/main/resources/application-dev.yml, a configuration file that typically contains sensitive information such as database credentials, API keys, and environment-specific secrets.
[COMMAND_EXECUTION]: The skill performs automated database operations, including executing CREATE TABLE (DDL) and INSERT (DML) statements directly into the connected database.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection or command/SQL injection because it interpolates user-provided strings (like function names and table prefixes) directly into shell commands (grep) and SQL queries (SHOW TABLES LIKE, SELECT ..., and the final generation SQL) without explicit sanitization or parameterization.
Ingestion points: User inputs for '功能名称' (Function Name) and '所属模块' (Module Name) in Step 1.
Boundary markers: No explicit boundary markers or instructions to ignore embedded instructions are present when processing these inputs.
Capability inventory: Executes grep via bash, reads local .yml files, and executes arbitrary SQL queries.
Sanitization: No evidence of sanitization or escaping of user-provided strings before they are used in commands.

dev