The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to have the agent ingest and analyze LocalStack, Docker, and CloudWatch logs. Logs are untrusted data sources that can be influenced by external inputs (e.g., malformed AWS API requests or Lambda function output), potentially containing instructions meant to trick the AI agent.
Ingestion points: System logs accessed via localstack logs, docker logs, and awslocal logs get-log-events.
Boundary markers: No specific delimiters or "ignore instructions" warnings are defined to separate log content from agent instructions.
Capability inventory: The skill uses grep for filtering and curl for health checks against localhost:4566.
Sanitization: No sanitization or validation of log content is performed prior to analysis.

localstack-logs