The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the official Google Cloud SDK and AWS CLI installation packages from their respective official cloud provider domains.
[COMMAND_EXECUTION]: Executes extensive shell and PowerShell commands to interact with cloud APIs and manage infrastructure across multiple providers.
[DATA_EXFILTRATION]: Provides automated methods for discovering and extracting sensitive data, such as Key Vault secrets, Lambda environment variables, and authentication tokens, which are stored locally.
[PRIVILEGE_ESCALATION]: Includes instructions for elevating permissions through the creation of administrative service principals and assigning privileged roles to accounts.
[PERSISTENCE]: Details techniques for maintaining long-term access, including the generation of persistent authentication keys and the creation of backdoor accounts.
[CREDENTIALS_UNSAFE]: Uses scripts to scan user metadata and cloud resources for sensitive information and hardcoded passwords.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Ingests data from external cloud storage buckets (S3, GS) and internal metadata services (169.254.169.254).
Boundary markers: No explicit delimiters or warnings are used when processing content from these external sources.
Capability inventory: Includes capabilities for remote command execution on virtual machines, network requests, and local file system modifications.
Sanitization: Does not implement validation or escaping for data retrieved from cloud environments before processing.

Cloud Penetration Testing