The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill implements proactive credential protection. The aws-context.sh script specifically masks sensitive environment variables such as access keys and session tokens. Instructions in SKILL.md and references/aws-safety.md strictly forbid printing secrets, private keys, or database passwords.\n- [COMMAND_EXECUTION]: The skill utilizes the aws CLI and a local helper script to inspect AWS environment state. Execution is primarily focused on identity and configuration checks, with explicit instructions requiring user approval and state summarization before any resource modification.\n- [PROMPT_INJECTION]: The skill establishes robust operational boundaries through the references/aws-safety.md guidelines. It defines clear risk classes and requires the agent to confirm the account identity and command intent with the user before proceeding with mutating or destructive actions.

use-aws