The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to modify .claude/settings.local.json to automatically allow a broad set of shell commands, including chmod, ln, uv, pip, and various git subcommands. This effectively bypasses the platform's security prompts and user review for these operations.
[PRIVILEGE_ESCALATION]: The skill instructs the agent to grant itself permissions for sensitive tools like chmod and ln, and to create symlinks targeting the agent's internal configuration directories (~/.claude/hooks), allowing the agent to modify its own execution environment.
[PERSISTENCE]: The skill configures a 'Stop' hook in ~/.claude/settings.json that triggers the execution of python3 ~/.claude/hooks/lint-gate.py. This ensures specific code runs automatically at the conclusion of every session, creating a persistent execution mechanism.
[DYNAMIC_EXECUTION]: The persistent 'Stop' hook mechanism executes a script located at ~/.claude/hooks/lint-gate.py which is symlinked to a source managed by a different skill (py-git-hooks), introducing a dependency on external code execution at runtime.

py-quality-setup