The Agent Skills Directory

[SAFE]: The skill is designed to guide an AI agent through the conceptual modeling of domain invariants, entities, and aggregates in Ruby projects. It operates entirely as a set of instructions and templates.
[PROMPT_INJECTION]: The skill includes behavioral constraints in the 'HARD-GATE' section (e.g., instructions to avoid implementation code unless requested). These are standard operational boundaries for the agent's task and do not attempt to bypass safety filters or override system instructions.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface by reading user-provided domain descriptions and requirements. However, it lacks dangerous capabilities—such as network exfiltration, arbitrary command execution, or sensitive file access—limiting the potential impact of adversarial content within the processed domain data.
Ingestion points: User descriptions of domain concepts (SKILL.md).
Boundary markers: Includes 'HARD-GATE' constraints to limit output scope.
Capability inventory: None (no shell execution, network ops, or file writes).
Sanitization: Not explicitly implemented, but risk is negated by lack of capabilities.
[METADATA_POISONING]: Metadata fields, including the version and origin, are consistent with the skill's purpose and match the provided author context ('igmarin').

model-domain