The Agent Skills Directory

Input Validation and Sanitization: The skill employs scripts/input_validator.py and specific validation logic in scripts/health_check.sh to verify user-supplied inputs such as instance names, VPC IDs, and port numbers. These checks ensure that parameters used in AWS CLI and shell commands adhere to expected patterns, reducing the risk of unintended command execution.
Secure Credential Management: The skill avoids hardcoded secrets by retrieving authentication tokens from AWS Secrets Manager using scripts/get_token.sh. These tokens are stored locally using restricted file permissions (chmod 600) to minimize exposure. The instructions also emphasize rotating initial passwords and using scoped API tokens for applications.
Indirect Prompt Injection Management: The skill processes requirements.json artifacts from related skills. To address potential indirect prompt injection, it implements a handoff contract that includes reading files from specific paths, validating them against a JSON schema, and performing scope checks before the data influences the agent's workflow.
TLS Verification Control: The scripts/health_check.sh utility includes an --insecure flag that allows disabling TLS certificate verification. The script includes a clear warning that this is intended for specific local port-forwarding scenarios and should not be used in production environments.
Least-Privilege Resource Operation: Guidance within the skill focuses on AWS resource mutations that are explicitly requested by the user. It recommends least-privilege IAM policies, private networking by default, and enforcement of encryption at rest using customer-managed KMS keys for production environments.

timestream-influxdb