The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the openskills read command to access content from other skills for integration purposes.
[EXTERNAL_DOWNLOADS]: The skill includes instructional templates for installing dependencies via npm, pnpm, or yarn as required by the integrated skills.
[PROMPT_INJECTION]: The skill ingests data from external skills which creates an indirect prompt injection surface. * Ingestion points: Content retrieved via openskills read in SKILL.md. * Boundary markers: No explicit delimiters or ignore-instructions warnings are used when processing the skill content. * Capability inventory: File-write access to the .trellis directory, subprocess execution for package managers, and command creation via /trellis:create-command. * Sanitization: No explicit sanitization of the external content is performed before interpolation into guidelines.

integrate-skill