The Agent Skills Directory

[PROMPT_INJECTION]: The reference file references/ch09-ai-serving.md contains the phrase 'ignore previous instructions'. Analysis confirms this is part of an educational explanation (Question 7) regarding RAG security and prompt injection risks, rather than a malicious instruction to the AI agent.- [COMMAND_EXECUTION]: The scripts/quiz.py script uses argparse and pathlib to provide a command-line interface for querying interview questions. It performs read-only operations on the bundled markdown files and does not invoke subshells or execute arbitrary strings.- [DATA_EXPOSURE]: While the skill discusses security topics such as JWT secrets and password hashing in references/ch06-security.md, it does not contain any hardcoded credentials, real secrets, or instructions to access sensitive system paths.- [EXTERNAL_DOWNLOADS]: The skill does not perform any network operations, use external APIs, or require third-party package installations. All logic is contained within the provided Python standard library script.

fastapi-ai-interview-prep