The Agent Skills Directory

[SAFE]: The skill does not include any scripts, shell commands, or executable code. It consists entirely of documentation and procedural instructions for backend resolution logic.
[INDIRECT_PROMPT_INJECTION]: The skill describes a resolution chain that involves reading local files such as .dh/config.yaml and .beads/dh-backend.
Ingestion points: .dh/config.yaml and the .beads/dh-backend marker file.
Boundary markers: None explicitly defined in the text.
Capability inventory: No scripts or tool calls are present in this skill.
Sanitization: The skill prescribes a specific resolution logic, which acts as a form of implicit validation by only looking for specific configuration keys or file existence.

backend-resolution