The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent and user to execute local scripts and batch files to process theme configurations, which involves running arbitrary code in the project environment.
Evidence: Step 4 in SKILL.md and the troubleshooting section in references/integration.md specify running node generate-theme.mjs and 生成并监听主题.bat.
[INDIRECT_PROMPT_INJECTION]: The skill processes theme.config files containing key-value pairs for theme settings. This represents an attack surface where malicious instructions could be embedded in the configuration data processed by the agent.
Ingestion points: theme.config (referenced in SKILL.md and provided as an example in assets/plum_mint/theme.config).
Boundary markers: None identified in the prompt templates or instructions.
Capability inventory: Shell command execution via node and file system write access for CSS generation.
Sanitization: No explicit sanitization or validation of the configuration file content is mentioned.
[DATA_EXFILTRATION]: The skill requires the agent to read and provide modifications for sensitive application configuration files to integrate themes, which could be exploited to expose or modify other sensitive settings.
Sensitive paths: Web.config (FineUI.Pro), appsettings.json (FineUI.Core), and Java Spring Boot components.
Evidence: The references/integration.md file provides specific snippets and instructions for modifying these sensitive files.

fineui-custom-theme