The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from existing workspace files to generate and modify workflow specifications.
Ingestion points: The skill reads from NOTES.md (user notes) and workflows/*.md (existing specifications) as defined in SKILL.md.
Boundary markers: There are no instructions providing delimiters or warnings to treat ingested content as data rather than instructions.
Capability inventory: The skill instructions authorize creating, editing, and deleting files within the workflows/ directory.
Sanitization: The skill does not define methods to sanitize or validate the content found in the ingestion points before processing it.

loop-me