The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines several bash-based validation routines using system utilities such as find, grep, sed, and jq. These commands are used to identify files and parse configuration data within the local plugin directory.- [PROMPT_INJECTION]: The skill performs validation on user-provided plugin files, which presents an indirect prompt injection surface. * Ingestion points: manifest files (plugin.json) and agent/command markdown files analyzed by the validator. * Boundary markers: None identified in the validation scripts. * Capability inventory: The skill utilizes subprocess calls to jq, python3, and bash for analysis tasks. * Sanitization: Shell snippets include basic quoting of variables to mitigate command injection in filenames.- [PROMPT_INJECTION]: An analysis of the skill's own instructions was performed to check for safety bypasses or system prompt extraction attempts. No override markers or jailbreak patterns were detected.- [EXTERNAL_DOWNLOADS]: The documentation contains a URL reference to the author's GitHub repository (github.com/poindexter12/waypoint) as a configuration example. No automatic download or remote execution logic is present.

plugin-validator