The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the omh command-line utility to perform diagnostic checks and environment repairs (e.g., omh doctor, omh setup). These operations are consistent with the skill's stated purpose of maintaining installation health.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied error messages and environment state as part of its diagnostic workflow. This represents an attack surface where maliciously crafted input could attempt to redirect the agent's logic.
Ingestion points: User-provided descriptions of issues and local environment metadata.
Boundary markers: The skill advocates for "explicit proof boundaries" when grouping diagnostic evidence, which provides some structural separation but does not fully isolate the agent from embedded instructions in the data.
Capability inventory: The skill is authorized to execute omh CLI commands and read local configuration files.
Sanitization: There are no explicit instructions for the agent to sanitize or escape user-provided issue descriptions before processing.

omh-doctor