The Agent Skills Directory

[SAFE]: The skill serves as a purely instructional guide for a well-known cloud service (Amazon Polly). It correctly identifies and documents API limits, engine capabilities, and operational boundaries based on official AWS documentation.
[DATA_EXFILTRATION]: No unauthorized data exfiltration patterns were detected. The skill actively promotes secure data handling by advising on IAM least-privilege roles, S3 bucket encryption (SSE-S3/SSE-KMS), and avoiding sensitive personal identifiers in free-form text fields that might be logged.
[REMOTE_CODE_EXECUTION]: The provided Python examples use the standard boto3 library for AWS interactions. No suspicious external dependencies, remote script downloads, or execution of untrusted code are present.
[PROMPT_INJECTION]: The skill does not contain instructions to override AI safety filters or ignore system prompts. It focuses on technical parameters for speech synthesis.
[INDIRECT_PROMPT_INJECTION]: While the skill involves processing external text and SSML (Speech Synthesis Markup Language) which could be untrusted, it provides necessary security guidance, including the requirement to escape reserved XML characters and awareness of AWS's internal safety mechanisms for generative models to prevent inappropriate output.
[PRIVILEGE_ESCALATION]: The skill follows the principle of least privilege, explicitly recommending that users grant only the specific IAM actions and S3 prefix permissions required for the task.

amazon-polly