Agent Research Files

Prompt Maker Skill

Research-backed principles for writing effective LLM prompts. Distilled from academic papers and Anthropic best practices. Focuses on permission-not-tutorial, context-first/task-last, positive framing with rationales, escape hatches, and prompt sensitivity testing.

HTML Output Skill

Rules for producing structured visual content as self-contained HTML files with light/dark toggle. Includes design system for font pairing, limited palette, cards over text walls, responsive layouts, and micro-interactions.

Research Files

Research.md

High-level agent systems research covering prompting, orchestration, tool-use, automation loops, and evaluation. The starting point for understanding the research landscape.

Research-prompt.md

Deep dive into single-agent prompting strategies: system prompt design, reasoning strategies, tool use, structured output, reliability, and uncertainty policy.

Research-orchestration.md

Agent orchestration and system design: workflows, multi-agent pipelines, memory, context management, error recovery, evaluation, and infrastructure patterns.