The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill is granted Bash(chmod:*) access, enabling it to modify file permissions. While justified by its purpose of validating script executability, this provides a mechanism for privilege modification.
[COMMAND_EXECUTION]: Executes shell-based commands using python3 and jq to perform schema validation and grading of plugin components.
[INDIRECT_PROMPT_INJECTION]: As the skill processes user-provided plugin directories, it is vulnerable to malicious instructions embedded in plugin files (e.g., SKILL.md or plugin.json) that could attempt to influence the validation output or the agent's behavior.

validate-plugin