The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from local concept YAML files and existing documentation to update or create guides, creating a surface where instructions embedded in these files could influence agent behavior.
Ingestion points: Content is read from plugins/runpod/skills/runpod-usage/concepts/*.yaml, SKILL.md, and reference/*.md.
Boundary markers: The instructions do not specify any delimiters or warnings to ignore embedded instructions within the processed data.
Capability inventory: The skill possesses the ability to execute local scripts via python3 and manage Node.js processes via pnpm.
Sanitization: No explicit sanitization, validation, or escaping of the ingested content is defined before it is processed or used to generate new documentation.
[COMMAND_EXECUTION]: The skill instructs the agent to execute various local commands and scripts for building, querying, and validating repository content.
Evidence: The skill calls pnpm build:sqlite, pnpm query, pnpm check:guides, and various Python hooks including hooks/check_links.py and hooks/check_runpod_branding.py. These represent functional capabilities for code execution within the repository's toolchain.

write-guide