The Agent Skills Directory

[COMMAND_EXECUTION]: The skill describes the commands section, which defines custom CLI commands that execute shell steps with argument and flag interpolation.
[EXTERNAL_DOWNLOADS]: The toolchain and vendor sections configure the downloading of external tools and components from remote registries (such as Aqua and Atmos) and repositories.
[PROMPT_INJECTION]: The skill documents a surface for indirect prompt injection where configuration manifests and CLI inputs are processed via template engines and executed as shell commands.
[PROMPT_INJECTION]: Ingestion points: atmos.yaml configuration, imported files, and command-line arguments.
[PROMPT_INJECTION]: Boundary markers: No delimiters or ignore instructions for embedded content are documented.
[PROMPT_INJECTION]: Capability inventory: Subprocess calls (via commands), file system access, and network operations (via auth, stores, and toolchain).
[PROMPT_INJECTION]: Sanitization: No explicit sanitization or validation of interpolated content is described in the documentation.

atmos-config