The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses bash to run a local script ${CLAUDE_PLUGIN_ROOT}/scripts/honne for persona checking and rendering tasks. These operations are scoped to the project's local directory and are standard for the skill's stated purpose.
[COMMAND_EXECUTION]: An inline python3 script is used to calculate the staleness of the persona data. This is a benign operation using standard libraries (json, datetime).
[DATA_EXPOSURE]: The skill reads and writes data within the .honne/ directory of the local project. Access is limited to project-relative paths, and no network exfiltration or access to sensitive system files is observed.
[PROMPT_INJECTION]: The skill processes data from .honne/persona.json to generate AI prompts. While this represents a potential surface for indirect prompt injection if the JSON contains untrusted content, the skill uses structured system prompt templates to manage the interpolation of this data safely.

persona