The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user data from the NOTES.md file to guide its interactive session and generate workflow specifications, creating a surface for indirect prompt injection.
Ingestion points: The skill explicitly reads the NOTES.md file located in the user's workspace to establish the context for the workflow design.
Boundary markers: No specific delimiters or instructions to ignore embedded instructions are defined for the content ingested from the NOTES.md file.
Capability inventory: The skill has access to powerful tools, including Bash, Write, and Edit, providing a potential pathway for follow-on actions if an injection were to occur.
Sanitization: The instructions do not specify any validation, filtering, or sanitization procedures for the text data imported from the workspace notes.

loop-me