The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides Python scripts to manage project-local memory files. All command execution is focused on these helper scripts (status.py, init.py, sortin.py, etc.) and does not involve executing arbitrary user-provided shell commands.
[DATA_EXFILTRATION]: No network operations or external data transmissions were detected. The skill only performs local file operations within the .ustht/ directory or imports data from specified project-local markdown files.
[PATH_TRAVERSAL]: The skill implements robust path validation in common.py (via ensure_runtime_path and validate_dim_name) to prevent directory traversal attacks. It explicitly checks for symlinks and ensures that all writes remain contained within the .ustht/ directory.
[DYNAMIC_EXECUTION]: The skill does not use eval(), exec(), or other dynamic code execution patterns. It processes markdown content as text data for organization and storage.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests user input and interpolates it into markdown files, it primarily serves as a storage mechanism. The sortin and resort workflows involve agent-led organization, but the skill includes safety guidelines to preserve original wording and avoid over-inference, which helps mitigate accidental obedience to embedded instructions.

user-thoughts