The Agent Skills Directory

[SAFE]: The skill follows security best practices for credential management, specifically recommending HashiCorp Vault or AWS Secrets Manager with KMS encryption instead of plaintext storage. It correctly uses environment variables for sensitive tokens (e.g., VAULT_TOKEN) and provides a legal compliance notice regarding the use of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill defines a data ingestion surface through the use of silo configuration files and deployment plans.
Ingestion points: contracts/schemas/deployment-plan.json is used to store silo topologies, asset IDs, and Vault paths for operational handoff.
Boundary markers: There are no explicit instructions for the agent to use XML tags or delimiters when parsing these structured data files to prevent embedded instruction execution.
Capability inventory: The skill is granted significant capabilities including execute_command, write_file, and edit_file across all files.
Sanitization: The instructions do not specify validation or sanitization procedures for the deployment-plan data, though the use of a JSON schema for output contracts provides some implicit structure.

manage-mmo-assets