The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local shell script (scripts/selfcheck.sh) to perform integrity checks on the installation. The script uses standard utilities like find, grep, and wc to analyze file existence and metadata within the skill's own directory structure.
[REMOTE_CODE_EXECUTION]: No remote code execution was detected. The skill instructions specifically forbid automatic networking, and the bash script contains no network-related commands (like curl or wget).
[DATA_EXPOSURE]: The skill reads local metadata from other skill files in the same package to ensure they are correctly installed. It does not access sensitive system files, environment variables, or user credentials.
[PROMPT_INJECTION]: The instructions provide clear boundaries for the diagnostic task and do not contain attempts to override safety filters or manipulate the agent's core behavior.

pm-selfcheck