The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by reading CloudWatch logs and CloudFormation events, which represents an indirect prompt injection surface.
Ingestion points: aws logs filter-log-events and aws cloudformation describe-stack-events are used to pull data into the agent's context.
Boundary markers: No specific delimiters are used to separate external log content from instructions.
Capability inventory: The agent has access to aws, sam, cdk, and terraform tools via bash.
Sanitization: The skill lacks specific instructions to sanitize or escape data retrieved from AWS logs.
[SAFE]: The allowed-tools field in the frontmatter correctly restricts the agent's environment to relevant infrastructure tools, minimizing the risk of unauthorized tool usage.
[SAFE]: The author 'aws-samples' provides standard, well-documented investigation commands that align with the skill's stated purpose of AWS debugging without any hidden or malicious behaviors.

aws-debug