prompt engineering

This entire page is part of IBM's prompt engineering guide and is structured around prompt engineering techniques.

Prompt engineering is the process of structuring inputs, and it has emerged as a crucial technique for maximizing the utility and accuracy of these models" — Direct definition and articulation of prom

clever prompts represented perhaps 0.1% of the total context modern AI systems process" — Article directly challenges prompt engineering as the primary driver of AI effectiveness in production. Contex

Context engineering combines prompt engineering, retrieval-augmented generation (RAG), and multi-agent techniques into one system, instead of using them separately." — Article explicitly names prompt

Over the past couple of years, building applications with large language models (LLMs) has shifted focus from prompt engineering to context engineering. In early LLM applications, users spent time cra

Author explicitly identifies prompt engineering as the turning point from failure to success in 2024→2025

Setting a role in the system prompt anchors tone and persona. Anthropic's prompt-engineering guidance is especially clear that role + clear instruction structure beats clever wording." — Article expli

Article explicitly positions context engineering as a replacement/displacement of prompt engineering, arguing that prompt tuning alone fails at scale for stateful systems with tools and agents.

The article demonstrates specific prompt design (adversarial framing) as a context engineering technique

2025: Ask AI better" — Article explicitly identifies prompting/prompt engineering as the 2025 competitive advantage in AI interaction

Context Engineering, a formal discipline that transcends simple prompt design to encompass the systematic optimization of information payloads for LLMs" — Survey explicitly positions context engineeri

Episode explicitly distinguishes context engineering from prompt engineering, treating them as separate disciplines with different focuses and skills

writing prompts (prompt engineering is just a small piece of it)" — Directly positions prompt engineering as subset of broader context engineering discipline, adding novel dimension that context encom

[direct] "I really like the term 'context engineering' over prompt engineering. It describes the core skill better: the art of providing all the context for the task to be plausibly solvable by the LL

GEPA extends static prompt engineering into dynamic, feedback-driven prompt evolution. Shifts from 'craft good prompt' to 'system that improves prompts iteratively'.

The shift from prompt engineering to Context Engineering is not just a matter of semantics; it's a response to the growing complexity of AI applications." — Article directly positions Context Engineer

When you decide to ask for structured output using XML tags, you are using an inference strategy. That inference strategy is independent of your task—it's about how you will render your prompt to show

Context engineering is building dynamic systems to provide the right information and tools in the right format such that the LLM can plausibly accomplish the task." — Article introduces 'context engin

Prompt engineering is key to getting llm agents to deliver predictable outcomes for test automation." — Article explicitly states that prompt engineering is critical for achieving reliable LLM agent b

I wrote a hook that explicitly forbids Claude from writing to and/or deleting those files" — Article demonstrates advanced constraint-based prompt engineering pattern where developer enforces strict r

The prompt was never the whole game." — Article directly challenges the premise that prompt engineering is the primary lever for AI system improvement, arguing context engineering is now the dominant

You basically create a python string with the prompt you want the LLM to process, and wherever you want to dynamically insert a variable like different product names, product categories, product subca

Article explicitly positions context engineering as evolution beyond simple prompt engineering. Karpathy quote distinguishes 'short task descriptions' (old prompt engineering) from 'delicate art of fi

Article explicitly positions context engineering as successor to/replacement for prompt engineering paradigm

Moves beyond single-prompt engineering to structured multi-layer context design; treats prompting as architecture problem

Article explicitly positions context engineering as the evolution PAST prompt engineering, arguing prompts were the 2022 bottleneck but context/goals are the 2026 bottleneck.

Explicitly distinguishes context engineering from prompt engineering, positioning it as a broader discipline focused on comprehensive context design vs. wording optimization

GPT-5.5 works best when prompts define the outcome and leave room for the model to choose an efficient solution path" — Article provides specific guidance on GPT-5.5 prompt optimization: shorter outco

Interaction with LLMs is facilitated through user and system prompts—carefully engineered instructions that ensure the model has all relevant information to perform the desired task accurately." — Art

Article explicitly distinguishes context engineering from prompt engineering, showing prompt engineering is insufficient for industrial applications and conflates short task descriptions with the full

Article explicitly positions context engineering as the evolution beyond prompt engineering, showing the structural additions that transform generic prompts into effective ones.

The seven prompting types (zero-shot, few-shot, CoT, etc.) are concrete instantiations of prompt engineering as a context engineering subdiscipline.

The core finding is that better prompting/harness design unlocked existing capabilities. This is applied prompt engineering.

A skill is a structured markdown file that teaches Claude Code how to perform a specific task on the Domino platform. Each skill contains context about relevant APIs, configuration patterns, and best

Article explicitly positions context engineering as superior to/different from prompt engineering for production systems, marking the transition from local optimization (prompts) to systemic optimizat

Article explicitly positions context engineering as replacement for prompt engineering paradigm, suggesting prompt-level optimization is insufficient

Positions context engineering as a distinct lever alongside prompt engineering, suggesting these are complementary disciplines

Length limits: keep text between tool calls to ≤25 words. Keep final responses to ≤100 words unless the task requires more detail." — Article demonstrates a specific system prompt constraint technique

Instead of modifying model weights, context adaptation improves performance after model training by incorporating clarified instructions, structured reasoning steps, or domain-specific input formats d

three specific changes to the "harness" surrounding the models had inadvertently hampered their performance" — Article directly documents how changes to system prompts and operational harnesses degrad

Article explicitly positions prompt engineering as a subset of context engineering—single-interaction optimization within a larger informational environment design problem.

developers use various prompt design strategies. The simplest involves providing clear instructions specifying a concrete action and result format" — Article directly demonstrates core prompt design s

Prompting isn't a magic trick. It's engineering." — Article explicitly frames prompt engineering as a software discipline, not an ad-hoc practice. Stephen Weber authored dedicated article on this topi

Context Engineering is the new name for prompt engineering. Success in RAG and AI agents is no longer about a single or simple prompt, it's about a complex sequence of inputs to the LLM" — Article exp

Despite our best efforts with prompt tuning, the model often hallucinated contacts, asked useless clarifying questions, or picked the wrong person because it was forced to process a massive amount of

Article's central thesis is that context engineering is distinct from and complementary to prompt engineering—prompt quality alone insufficient

Article explicitly positions context engineering as the evolution beyond prompt engineering, showing that prompt engineering is narrower (instructions only) while context engineering encompasses the f

What the agent is and what it should always or never do" — Directly articulates core function of system prompt in defining agent identity and constraints

Context engineering has replaced prompt engineering as the main challenge in building agents and LLM applications" — Article explicitly positions context engineering as an evolution/replacement of pro

Don't use prompts for control flow. Use control flow for control flow." — Article explicitly argues against using prompts as a control mechanism, advocating for actual control flow structures instead.