The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional text guiding the agent on how to specialize its persona and output for different industries. It does not perform any network operations, file system modifications, or execution of external code.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided domain information to tailor prompts, it defines an instructional surface rather than a malicious one. It specifically encourages the inclusion of safety guardrails and disclaimers (e.g., 'Not legal advice') when specializing, which serves as a defensive measure.
Ingestion points: [domain] argument defined in the frontmatter of SKILL.md.
Boundary markers: None present for the [domain] input.
Capability inventory: No dangerous capabilities (file access, network, or command execution) are present in the skill.
Sanitization: No explicit sanitization of the [domain] input is described, but the skill's purpose is to add safety constraints.

specialize