Prompt Engineering Guide
MIT-licensed comprehensive guide and prompt library — techniques, examples, and templates for every major LLM prompting method.
Prompt Engineering Guide
The Prompt Engineering Guide by DAIR.AI is the most comprehensive open-source reference for LLM prompting — covering zero-shot, few-shot, chain-of-thought, ReAct, tree-of-thought, self-consistency, generated knowledge, and more. It includes model-specific guidance, prompt templates, research paper summaries, and a growing library of community-contributed examples.
Key features
- Complete technique coverage: zero-shot, few-shot, CoT, ReAct, PAL, tree-of-thought, RAG prompting
- Model-specific notes for GPT-4, Claude, Gemini, Mistral, Llama, and others
- Prompt templates for code generation, summarization, classification, extraction
- Research-backed: cites and explains 100+ prompting papers
- Available as a website, PDF, and raw Markdown
- MIT license — copy, adapt, integrate commercially
- Available in 13 languages
Key prompt patterns
# Chain-of-Thought (CoT) — few-shot example
Q: Roger has 5 tennis balls. He buys 2 more cans, each with 3 balls. How many?
A: Roger started with 5 balls. 2 cans × 3 balls = 6 more. 5 + 6 = 11. The answer is 11.
Q: The cafeteria had 23 apples. They used 20 for lunch and bought 6 more. How many?
A: [Let the model reason step-by-step here]
# ReAct pattern
Thought: I need to find the current population of Tokyo.
Action: Search[Tokyo population 2024]
Observation: Tokyo's population is approximately 13.96 million.
Thought: I have the answer.
Answer: Tokyo's population is approximately 13.96 million.
Install via ai-supply
npx ai-supply add prompt-engineering-guide
Curated mirror of the open-source Prompt Engineering Guide (MIT). Get it from the source.
Compromise signals — malicious or tampered code (leaked secrets, backdoors, a dropped executable) — reduce the score, and known dependency CVEs carry a bounded penalty (they warrant review but never QUARANTINE — update the dependency to clear). Other dangerous-by-capability traits are risk surface, expected for some capabilities. Every finding is mapped to its OWASP control below.
Findings mapped to the OWASP Top 10 for LLM Applications (2025) and the OWASP Machine Learning Security Top 10. Expand any flagged control for the exact findings — compromise reduces the score; expected/risk-surface do not, except a known CVE, which carries a small bounded penalty (high/critical → Review).
The same gate an agent runs before installing (POST /api/v1/trust/prompt-engineering-guide/check). Click a policy:
Consume Prompt Engineering Guide programmatically. Authenticate with an API key or session — see Authorize an agent.
# Agents: CHECK BEFORE YOU INSTALL (no auth) — score, grade, level, capability manifest
curl https://ai-supply.store/api/v1/trust/prompt-engineering-guide
# Gate against your org policy (returns { pass, violations })
curl -X POST https://ai-supply.store/api/v1/trust/prompt-engineering-guide/check \
-H "Content-Type: application/json" \
-d '{"minGrade":"B","denyPermissions":["shell"],"denyUnknownEgress":true}'
# CLI
npx ai-supply add prompt-engineering-guide
# REST (install → download)
curl -X POST https://ai-supply.store/api/v1/listings/prompt-engineering-guide/install \
-H "Authorization: Bearer $AIM_KEY"
# MCP tool
install_listing({ "slug": "prompt-engineering-guide" })OpenAPI spec →Curated mirror — latest upstream source. See the repository for tagged releases.