The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents standard AWS CLI usage for SNS, SQS, MQ, and Step Functions, which are intended for resource management and administrative tasks within the AWS environment.- [REMOTE_CODE_EXECUTION]: The skill uses the standard boto3 library for Python to interact with AWS services, which is the official and expected tool for the skill's stated purpose.- [DATA_EXFILTRATION]: The skill provides methods to retrieve metadata and message contents from AWS services (e.g., sqs.receive_message). While this involves handling potentially sensitive data, it is limited to official AWS endpoints and no patterns of exfiltration to unauthorized domains were detected.- [PROMPT_INJECTION]: The skill exposes a potential attack surface for indirect prompt injection via the retrieval of message content from SQS queues.
Ingestion points: aws sqs receive-message and sqs.receive_message() in SKILL.md.
Boundary markers: No specific delimiters or instructions to ignore external instructions are present in the snippets.
Capability inventory: The agent has the capability to publish messages and start workflow executions, which could be influenced by malicious content in ingested messages.
Sanitization: No explicit content validation or sanitization is implemented in the provided examples.

aws-messaging