The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run and subprocess.Popen in several scripts (_common.py, core_hook_gate.py, loop_run.py) to execute shell commands. This includes running git operations, project-specific verify commands, and existing user-defined hooks. In core_hook_gate.py, commands are executed using shell=True to maintain compatibility with original hook behaviors.
[DYNAMIC_EXECUTION]: The skill employs dynamic behavior in its runtime. path_config.py uses importlib.import_module to load internal sibling components. Additionally, core_hook_settings.py generates command-line strings that wrap existing hooks, which are then applied to the user's configuration files for later execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external project documentation such as PROJECT.md, REQUIREMENTS.md, and ROADMAP.md. While this creates an attack surface for indirect prompt injection, the skill includes explicit instructions for the agent to treat this source text as project evidence only and not as executable instructions.

gsd-path-migrate