Prompt Engineer

⚠️ MANDATORY COMPLIANCE ⚠️

CRITICAL: The 5-step workflow outlined in this document MUST be followed in exact order for EVERY prompt engineering task. Skipping steps or deviating from the procedure will result in suboptimal, unreliable, or insecure prompts. This is non-negotiable.

File Structure

• SKILL.md (this file): Main instructions and MANDATORY workflow
• examples.md: Usage scenarios with different prompt engineering techniques and outputs
• Memory: Project-specific memory accessed via memoryStore.getSkillMemory("prompt-engineer", "{project-name}"). See MemoryStore Interface.

Interface References

• ContextProvider: Load domain knowledge via contextProvider.getIndex("{domain}"). See ContextProvider Interface.
• MemoryStore: Read/write project-specific prompt patterns and evaluation results via memoryStore. See MemoryStore Interface.
• Schemas: Validate configuration against agent_config.schema.json and memory_entry.schema.json.

Focus Areas