The Agent Skills Directory

[Indirect Prompt Injection Surface]: The skill ingests workload requirement artifacts from external JSON files to inform database recommendations. While the skill includes steps for schema validation and scope checking, processing external data is a potential injection vector that could influence agent behavior or recommendations.
[Credential Management]: The get_token.sh script retrieves API tokens from AWS Secrets Manager and stores them in a local hidden file (~/.influxdb_token). The script implements security measures such as umask and chmod to restrict file access to the current user, which is a common pattern for local credential caching in development environments.
[Network Connectivity Verification]: The health_check.sh script performs outbound HTTPS requests to verify instance availability based on user-provided endpoints. While this is a standard diagnostic function, it involves interaction with network resources and includes an optional flag to bypass TLS verification, which is noted as a security consideration for non-production environments.
[Command Execution Environment]: The skill provides several shell scripts for VPC endpoint validation and health checks. These scripts utilize standard system tools like curl, aws CLI, and python3 for processing. They include input validation patterns to mitigate command injection risks when handling user-provided parameters like endpoints and port numbers.

timestream-influxdb