The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation provides a shell command in section 13 for persistent installation within the Claude Code environment. This command modifies the ~/.claude/CLAUDE.md file to include the skill's URL, which is a standard procedure for that platform.
[EXTERNAL_DOWNLOADS]: The skill references medical clinical guidelines from trusted sources such as the CDC, WHO, and NICE for evidence-based practice.
[PROMPT_INJECTION]: The skill is intended to process external patient data and case histories (Category 8 surface). While it functions as a reasoning framework, it is noted that external data ingestion enters the context at file and prompt level (§12). No explicit boundary markers or sanitization are defined, but the risk is low given the specific clinical context.
[PROMPT_INJECTION]: The skill uses identity definitions to establish a medical persona. No patterns for bypassing safety guidelines or overriding system instructions were detected.

attending-physician