Comprehensive prompt and context engineering. Every recommendation grounded in research.
| system prompt | The top-level instruction block sent before user messages; sets model behavior | | context window | The full token budget: system prompt + conversation history + tool results + retrieved docs | | context engineering | Designing the entire context window, not just the prompt text — write, select, compress, isolate |
| template | A reusable prompt structure with variable slots ({{input}}, $ARGUMENTS) | | rubric | A scoring framework with dimensions, levels (1-5), and concrete examples per level | | few-shot example | An input/output pair included in the prompt to demonstrate desired behavior |
Comprehensive prompt and context engineering for any AI system. Four modes: (1) Craft new prompts from scratch, (2) Analyze existing prompts with diagnostic scoring and optional improvement, (3) Convert prompts between model families (Claude/GPT/Gemini/Llama), (4) Evaluate prompts with test suites and rubrics. Adapts all recommendations to model class (instruction-following vs reasoning). Validates findings against current documentation. Use for system prompts, agent prompts, RAG pipelines, tool definitions, or any LLM context design. NOT for running prompts, generating content, or building agents. Source: wyattowalsh/agents.