The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard shell commands (mkdir, echo) to create project directories and configure the .gitignore file, which is appropriate for its initialization purpose.
[REMOTE_CODE_EXECUTION]: It defines a validation hook that executes a local Python script (writing-precis-guard.py) from the plugin's internal directory to ensure artifact quality.
[PROMPT_INJECTION]: The skill exhibits a surface area for indirect prompt injection by reading session state from .planning/HANDOFF.md and accepting user-provided answers for the writing templates.
Ingestion points: Content is read from .planning/HANDOFF.md and from AskUserQuestion tool responses.
Boundary markers: Absent; the skill reads and displays handoff content directly without applying delimiters or instructions to ignore embedded commands.
Capability inventory: File system operations (read/write), shell command execution, and local script execution via hooks.
Sanitization: Absent; the skill does not sanitize or escape external data before it is processed or written to files.

writing-setup