The Agent Skills Directory

[COMMAND_EXECUTION]: The diagnostic script scripts/director-doctor.py uses subprocess.run to execute git commands and verify the installation status of CLI tools (claude, codex, grok) by checking their version information.
[DATA_EXFILTRATION]: The scripts/director-doctor.py script accesses and reads configuration files within user-home directories, such as ~/.claude/settings.json, ~/.codex/hooks.json, and ~/.grok/hooks.json, to report on hook registrations and overall system health for Director Mode.
[PROMPT_INJECTION]: The skill features a surface for indirect prompt injection by processing repository-level configuration data that is then reported to the agent.
Ingestion points: The script scripts/director-doctor.py reads data from JSON configuration files including .claude/settings.json, .codex/hooks.json, and .grok/hooks.json.
Boundary markers: Absent; although data is parsed as JSON, it is not encapsulated in specific protective delimiters when presented in the diagnostic report.
Capability inventory: The script performs diagnostic subprocess.run calls to verify system tools and versions.
Sanitization: The script performs basic JSON structure and data type validation during the reporting process.

director-mode