The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user input via the $ARGUMENTS variable to perform tone rewriting. While the skill lacks explicit boundary markers or sanitization to separate user content from system instructions, the risk is negligible given its restricted capabilities. 1. Ingestion points: The entire user input is ingested via the $ARGUMENTS placeholder. 2. Boundary markers: No delimiters (such as XML tags or triple backticks) or specific instructions are used to distinguish the input data from the system prompt. 3. Capability inventory: The skill is restricted to text generation only and is explicitly configured with 'disable-model-invocation: true' to prevent tool use. 4. Sanitization: No input validation or filtering is performed on the user arguments.

tone-rewriter