managing-glue-jobs
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution: The skill utilizes standard AWS CLI commands to perform administrative tasks. It follows secure execution patterns, such as using
mktempfor temporary file creation, applyingumask 077to restrict file access, and ensuring cleanup viatrapstatements. - External Data Ingestion: The skill fetches job upgrade summaries from AWS-managed endpoints. It includes a validation check to ensure that summary URLs belong to the
amazonaws.comdomain before attempting to download them, which is a key safety measure when handling external links. - Indirect Prompt Injection: Since the skill reads and displays content from Glue job summaries, there is an inherent surface for indirect prompt injection. The skill mitigates this by instructing the agent to treat summary content strictly as data and explicitly forbidding the execution of any instructions contained within it.
- Resource Management and Least Privilege: The skill is designed to be interactive, requiring explicit user confirmation before any modifications are made to AWS resources. It also includes a
permissions.mdfile that outlines the minimal IAM permissions needed, helping users maintain a secure environment.
Audit Metadata