The Agent Skills Directory

[SAFE]: The skill facilitates a design session aimed at producing workflow specifications in the local workspace. It operates exclusively through interactive dialogue and management of local markdown files in the workflows/ directory and a NOTES.md file. No high-risk behaviors such as network communication, command execution, or sensitive data access were identified.- [INDIRECT_PROMPT_INJECTION]: The skill ingests user context from local files and conversation to generate specifications.
Ingestion points: Reads user world information from NOTES.md and user conversation.
Boundary markers: The instructions do not define explicit delimiters for ingested content.
Capability inventory: Limited to creating, editing, and deleting markdown files within the specific workflows/ directory. No network, subprocess, or dynamic code execution capabilities are defined.
Sanitization: No specific content sanitization or validation is mentioned. Given the lack of dangerous capabilities, the risk associated with processing this data is safe.

loop-me