The Agent Skills Directory

[SAFE]: The skill's primary logic involves creating and maintaining a stateful decision map in a local markdown file. No malicious commands, obfuscated content, or unauthorized network access patterns were detected.
[PROMPT_INJECTION]: The skill processes user-generated markdown files ('decision maps') as context, which presents an indirect prompt injection surface. This behavior is consistent with the skill's core purpose of tracking project-specific decisions across sessions. Ingestion points: local project directory markdown files. Boundary markers: none identified in instructions. Capability inventory: coordinates with other agent skills like /prototype and /grilling. Sanitization: none specified. The risk is considered inherent to the project-management workflow and is assessed as safe.

decision-mapping