The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze local diagnostic data, such as installation logs, managed skill metadata, and system configuration files. This creates a vulnerability surface where maliciously crafted data within these external files could attempt to influence the agent's diagnostic conclusions.
Ingestion points: Local managed skills directories, Hermes config files, and the user-provided observed issue description.
Boundary markers: The skill explicitly instructs the agent to use "explicit proof boundaries" and to avoid collapsing different diagnostic claims into a single verification.
Capability inventory: The skill uses the omh CLI for recording diagnostics and performing environment repairs (update/setup).
Sanitization: The skill emphasizes evidence-gated reporting and verification but does not detail specific character-level sanitization for log content.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute specific local commands, such as omh runtime record, to document diagnostic sessions. These commands are restricted to the environment's management tool and are appropriate for the skill's stated purpose as a diagnostic 'doctor' workflow.

omh-doctor