The Agent Skills Directory

[SAFE]: The skill consists entirely of architectural principles, glossaries, and workflow documentation. No obfuscated code, hidden URLs, unauthorized network operations, or malicious persistence mechanisms were identified in any of the provided files.
[PROMPT_INJECTION]: The 'Design It Twice' pattern in DESIGN-IT-TWICE.md describes a workflow that spawns sub-agents and provides them with technical briefs derived from user-provided codebase information. This establishes a surface for indirect prompt injection.
Ingestion points: User-provided technical briefs, file paths, and coupling details extracted from the local codebase.
Boundary markers: The skill does not explicitly instruct the agent to use delimiters or warnings for the sub-agents to ignore instructions embedded within the codebase content.
Capability inventory: The sub-agents are strictly tasked with architectural design brainstorming (generating interfaces and examples) and are not provided with capabilities for system commands, file modifications, or network access.
Sanitization: There are no instructions for sanitizing or filtering input derived from the codebase before it is passed to the sub-agents.

codebase-design