The Agent Skills Directory

[SAFE]: The skill is entirely informational, providing guidance on prompt engineering patterns such as Few-Shot Learning, Chain-of-Thought, and Template Systems. It does not contain any executable scripts or automation logic that could pose a threat.
[PROMPT_INJECTION]: While the skill discusses 'Persuasion Principles' (e.g., Authority, Commitment) and the use of imperative language like 'YOU MUST' to ensure compliance, these are framed as educational techniques for improving LLM reliability in development tasks. There are no attempts to bypass safety filters or hijack the host agent's instructions.
[REMOTE_CODE_EXECUTION]: The document includes illustrative code snippets for Python and Shell (e.g., using pdfplumber or running a migration script). These are static examples provided for user reference and do not involve the agent downloading or executing code from remote or untrusted sources.

prompt-engineering