The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands (bash, find, grep, sed, cmp) to perform environment discovery and verify the integrity of project artifacts. It resolves paths from the local plugin cache and compares project-level files against their original assets to detect manual edits or stale installations.
[PROMPT_INJECTION]: The skill processes untrusted metadata (such as version strings and ownership tags) from files within the project directory to populate its status report. While this presents an indirect prompt injection surface, the risk is considered safe due to the strict formatting requirements for these fields and the skill's read-only operation.
[SAFE]: The skill implements safety best practices by explicitly refusing to automate interactive setup processes that could result in unexpected side effects. It provides a structured 'PLAN' block and insists on manual user execution of recommended commands.

setup-status