The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a suite of shell scripts in the scripts/ directory designed for project perception. These scripts use standard tools like rg, grep, ls, cat, and awk to search the project's Java/Kotlin source code for API-related patterns and to read configuration files. These actions are restricted to the local environment and are used to provide the agent with necessary context for rule authoring.
[DATA_EXPOSURE]: Helper scripts (list_rule_files.sh, read_rule_file.sh) are designed to access the user's ~/.easyapi directory to read global configuration files. This is a functional requirement for the skill to understand global rule precedence and avoids the use of arbitrary file access tools.
[PROMPT_INJECTION]: The skill's core instructions and bundled recipe catalog include explicit 'Critical Quality Rules' that instruct the agent to never hardcode credentials and to avoid generating broad field-ignore rules for sensitive data unless specifically requested. These guidelines mitigate the risk of the assistant inadvertently suggesting insecure configurations.
[REMOTE_CODE_EXECUTION]: While the skill assists in writing Groovy and JavaScript snippets for the EasyApi plugin, it specifically notes that these scripts are intended to run within the plugin's sandboxed environment. The skill itself does not facilitate the execution of remote or untrusted code on the local machine.

easy-api-assistant