The Agent Skills Directory

[SAFE]: The skill follows a structured methodology for software design and implementation. It utilizes designated memory tools (mcp__claude-flow__memory_*) for state management and standard file manipulation tools (Read, Write, Edit, Bash) for code production.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves specifications from a memory search which are then used to inform pseudocode and architectural designs. While this represents a potential surface for indirect prompt injection if the specifications are derived from untrusted external data (like public PR comments), it is a standard operational pattern for AI coding assistants.
Ingestion points: Specifications retrieved via mcp__claude-flow__memory_search in SKILL.md.
Boundary markers: Not explicitly defined in instructions.
Capability inventory: Includes Bash, Write, Edit, and mcp__claude-flow__workflow_create capabilities.
Sanitization: Not explicitly implemented in the prompt instructions.

sparc-implement