The Agent Skills Directory

[NO_CODE]: The skill consists entirely of markdown instructions and does not include any executable scripts, binaries, or platform configuration files.
[COMMAND_EXECUTION]: The documentation provides examples for using the meridian command-line tool. These are illustrative examples for the agent to follow and do not represent autonomous or hidden command execution.
[DATA_EXFILTRATION]: The instructions involve reading from local workspace directories (e.g., $MERIDIAN_FS_DIR/characters/, $MERIDIAN_WORK_DIR/outline/). These operations are scoped to the agent's project environment and are intended for retrieving story context rather than harvesting sensitive system data.
[PROMPT_INJECTION]: The skill defines a workflow for ingesting external content (prior chapters and outlines), which constitutes a potential surface for indirect prompt injection. However, the skill provides these as guidelines for standard operation and does not contain malicious injection payloads itself.

story-context