The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted Product Requirements Documents (PRDs) to generate implementation milestones and verification commands, which creates an indirect prompt injection surface where a malicious document could influence output commands. * Ingestion points: Extraction of behavior and constraints from external PRD files as defined in SKILL.md. * Boundary markers: None identified in the instructional workflow. * Capability inventory: Analysis of codebase paths and generation of executable verification commands. * Sanitization: No explicit logic for filtering or escaping external requirements content is provided.

to-milestones