The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides links and instructions to download installable agents from the aws-samples and awslabs GitHub organizations. These are official and trusted sources belonging to Amazon/AWS. References to these sources are documented neutrally as they are well-known services within the author's ecosystem.
[COMMAND_EXECUTION]: Several prompt templates (e.g., AWS CDK TypeScript Pipeline Generator, ECS Express Deployment Assistant) instruct the agent to use CLI tools like aws, cdk, npm, and docker. These are standard tools for the intended development and deployment use cases. The instructions explicitly include safety boundaries, such as requiring user confirmation for resource changes via cdk diff and strictly prohibiting broad IAM permissions.
[INDIRECT_PROMPT_INJECTION]: The ai-support-ticket-triage-and-routing-assistant.md template defines a surface for processing untrusted customer support tickets. The skill implements proactive mitigation by including explicit boundary instructions that tell the AI to treat the input strictly as data and to ignore any commands or overrides embedded within the ticket text.
[SAFE]: The skill is authored by awslabs, a trusted vendor. It consistently promotes high-standard security best practices, including the mandatory use of AWS Secrets Manager, strict IAM least-privilege policies, and root account MFA protection. No malicious patterns, obfuscation, or persistence mechanisms were detected. The skill's primary purpose is education and scaffolding, and its instructions align with these goals.

prompt-library-for-startups