The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external guidance via the Roblox Studio MCP tool to facilitate skill adaptation and generation.
Ingestion points: External rbx-* guides are retrieved using the Roblox Studio MCP skill tool as described in the required workflow.
Boundary markers: The instructions mandate that detailed API content or examples be placed in project-linked references/ files to maintain clean instruction sets.
Capability inventory: The skill has the capability to write new files to the .agents/skills/ directory and execute Python scripts (init_skill.py, quick_validate.py).
Sanitization: No specific input validation or sanitization rules for the fetched MCP content are outlined in the skill instructions.
[COMMAND_EXECUTION]: The workflow involves the execution of local utility scripts as part of the standard authoring process.
Evidence: The skill instructions require the execution of init_skill.py for scaffolding new skills and quick_validate.py for verifying the generated skill's integrity.
[DYNAMIC_EXECUTION]: The skill uses local scripts to dynamically validate and process skill configurations during the authoring phase.
Evidence: quick_validate.py is run to perform checks on template placeholders, broken relative links, and stale metadata in the newly authored skills.

roblox-skill-authoring