The Agent Skills Directory

[PROMPT_INJECTION]: The CliPlugin documentation includes an example that processes data from a ZModel AST and writes it to the file system. This pattern represents a standard vulnerability surface for indirect prompt injection where data from a schema file influences generated artifacts.\n
Ingestion points: The generate callback in SKILL.md accepts the model object (AST) representing the ZModel schema.\n
Boundary markers: The code example does not use explicit boundary markers or delimiters for the schema-derived data.\n
Capability inventory: The example demonstrates the use of fs.writeFileSync to write generated content to the local disk.\n
Sanitization: The example does not include explicit sanitization of the AST content before it is written to the file system, which is standard for illustrative development documentation.\n- [EXTERNAL_DOWNLOADS]: The skill recommends installing @zenstackhq/sdk and references @zenstackhq/orm, which are the official development packages provided by the author.\n- [COMMAND_EXECUTION]: The provided code snippets demonstrate expected developer capabilities such as writing files to an output directory and mention the framework's use of jiti for loading TypeScript modules.

zenstack-plugin-dev