The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute a local Python script (scripts/sync_codex_agents.py) to perform its primary function of synchronizing configuration files.
Evidence: SKILL.md contains bash code blocks such as python3 scripts/sync_codex_agents.py --scope global intended for manual or agent-led execution.
[SAFE]: The script performs file system operations (reading templates and writing TOML configurations) within expected and documented paths (~/.codex/agents/ and .codex/agents/).
Evidence: scripts/sync_codex_agents.py uses Path.home() and repo-relative paths to determine target directories.
[SAFE]: The script implements a safety mechanism to prevent accidental overwriting of files not managed by the tool.
Evidence: MANAGED_HEADER is used in scripts/sync_codex_agents.py to identify and protect unmanaged files unless the --force flag is explicitly provided.
[SAFE]: The identified subprocess.run call in the test suite is used strictly for integration testing of the local synchronization script.
Evidence: scripts/test_sync_codex_agents.py invokes subprocess.run to execute the sibling script scripts/sync_codex_agents.py in a controlled test environment.

codex-superra-setup