The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill includes a Python script (scripts/calculate_metrics.py) and onboarding documentation (references/day1_onboarding.md) that involve command execution.
The Python script is a utility for processing QA data from a CSV file.
The onboarding guide contains standard development commands such as git clone, pnpm install, and docker exec for setting up a testing environment.
[EXTERNAL_DOWNLOADS] (LOW): Documentation references the installation of external dependencies and repository cloning using standard package managers. This is documented as part of a manual setup process rather than an automated, hidden runtime execution.
[INDIRECT_PROMPT_INJECTION] (INFO): The metrics script processes external CSV data. However, the logic is limited to calculating and displaying numeric counts and does not involve re-injecting untrusted data into an LLM prompt in a way that could influence agent behavior.
[DATA_EXPOSURE] (INFO): Documentation mentions paths like ~/.claude/skills/ and .env files. These are referenced in the context of standard CLI tool configuration and local development setup, with no code present to exfiltrate these files to external domains.

qa-expert