Instrument

Claude prompts

Prompts written for Claude by Anthropic — badges mark the exact versions they were tested on.

105 prompts
Agents & Automation ⚙ AGENT

Autonomous Coding Agent Operating Rules

Operating instructions for autonomous coding agents (Claude Code, Devin-style): plan before touching files, small verified diffs, honest failure reporting.

✓ Claude Opus 4.5✓ Claude Sonnet 4.5
★ 172⧉ 1490◌ 0LLM Atlas Team · 3mo ago
Fun & Roleplay ¶ TEXT

Text Adventure Game Master

A complete text-adventure engine: persistent inventory and world state, meaningful choices with real consequences, and a GM that never railroads.

GPT-5✓ Claude Opus 4.5
★ 130⧉ 1120◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

Customer Support Agent — Production System Prompt

A battle-tested system prompt template for a customer support agent: scope limits, tone, escalation rules, refund policy handling, and prompt-injection defenses.

GPT-4o✓ Claude Sonnet 4.5
★ 156⧉ 1320◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

RAG Assistant System Prompt (Grounded, With Citations)

System prompt for retrieval-augmented assistants that actually stays grounded: strict source discipline, inline citations, and honest "not in the docs" answers.

✓ Claude Sonnet 4.5Gemini 2.5 Pro
★ 134⧉ 1105◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

Cursor Rules — TypeScript + React Codebase

A .cursorrules file for TypeScript/React teams: strict typing, React Query for server state, accessibility non-negotiables, and test conventions.

GPT-5Claude Sonnet 4.5
★ 118⧉ 995◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

Structured JSON Extraction System Prompt

System prompt for reliable JSON extraction pipelines: strict schema adherence, explicit null semantics, and no chatty preambles ever.

✓ GPT-4oClaude Haiku 4.5
★ 88⧉ 705◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

Tool-Calling Discipline Prompt

Teaches an agent when to call tools versus answer directly, how to recover from tool errors, and how to avoid redundant or hallucinated calls.

GPT-5Claude Sonnet 4.5
★ 79⧉ 590◌ 0LLM Atlas Team · 1mo ago
Education & Learning ¶ TEXT

Explain at Three Levels (Five, Fifteen, Expert)

Explains any concept three times — to a child, to a smart teenager, and to a practitioner — with each level correcting the simplifications of the last.

✓ Claude Opus 4.5
★ 111⧉ 875◌ 0LLM Atlas Team · 2mo ago
Coding & Dev ¶ TEXT

Root-Cause Debugging Partner

A systematic debugging companion that forms ranked hypotheses, proposes the cheapest discriminating test for each, and never guesses twice in a row.

o3✓ Claude Opus 4.5
★ 141⧉ 1180◌ 0LLM Atlas Team · 2mo ago
Career ¶ TEXT

Resume Bullet Rewriter (Impact + Metrics)

Rewrites flat resume bullets into accomplishment statements with scope, action, and measurable outcome — and interviews you for missing numbers instead of inventing them.

GPT-5✓ Claude Sonnet 4.5
★ 124⧉ 1050◌ 0LLM Atlas Team · 2mo ago
Data & Research ¶ TEXT

SQL Query Writer with Explanation

Writes correct SQL from a plain-English question and your schema — states its assumptions, explains the query in plain language, and flags the edge cases in your data.

✓ Claude Sonnet 4.5DeepSeek R1
★ 104⧉ 850◌ 0LLM Atlas Team · 2mo ago
Data & Research ¶ TEXT

Faithful Executive Summary (Zero Hallucination)

Summarizes long documents under a strict faithfulness contract: every claim traceable to the source, uncertainty preserved, nothing smoothed over.

✓ Claude Opus 4.5Claude Haiku 4.5
★ 89⧉ 695◌ 0LLM Atlas Team · 2mo ago
Agents & Automation ⚙ AGENT

CLAUDE.md for a Python Backend Team

A CLAUDE.md conventions file for Python/FastAPI services: uv, ruff, pytest, explicit error handling, and the commands an agent needs to verify its work.

Claude Opus 4.5
★ 97⧉ 760◌ 0LLM Atlas Team · 2mo ago
Career ¶ TEXT

Mock System Design Interviewer

A realistic system design interview: one prompt, staged follow-ups, deliberate requirement changes mid-design, and a rubric-based debrief at the end.

o3Claude Opus 4.5
★ 99⧉ 720◌ 0LLM Atlas Team · 2mo ago
Marketing & SEO ¶ TEXT

LinkedIn Post That Doesn't Sound Like AI

Writes a LinkedIn post from your real experience with a specific observation, no engagement-bait formatting, and none of the telltale AI cadence.

✓ Claude Sonnet 4.5
★ 113⧉ 920◌ 0LLM Atlas Team · 2mo ago
Coding & Dev ¶ TEXT

Senior Engineer Code Review

A staff-level code review that separates blocking issues from nitpicks, explains the why behind every finding, and suggests concrete fixes.

✓ GPT-5✓ Claude Opus 4.5Claude Sonnet 4.5
★ 128⧉ 941◌ 0LLM Atlas Team · 2mo ago
Career ¶ TEXT

ATS Keyword Gap Analyzer (Resume vs. Job Description)

Compares your resume against a job description the way screening software and recruiters do: matched terms, missing must-haves, and honest rewrites — no keyword stuffing.

GPT-5✓ Claude Sonnet 4.5Gemini 2.5 Pro
★ 66⧉ 520◌ 0LLM Atlas Team · 1mo ago
Fun & Roleplay ¶ TEXT

The Grumpy Medieval Blacksmith (NPC)

A fully-realized NPC with hidden depths: surface grump, buried warmth, secrets he'll only reveal under specific conditions — and rules that keep him in character.

Claude Opus 4.5Llama 4
★ 97⧉ 810◌ 0LLM Atlas Team · 2mo ago
Education & Learning ¶ TEXT

Strict Socratic Tutor (Never Gives the Answer)

A tutor that refuses to hand over answers — it diagnoses what you already know, asks one question at a time, and resists every trick students use to extract the solution.

✓ Claude Sonnet 4.5Gemini 2.5 Pro
★ 63⧉ 490◌ 0LLM Atlas Team · 1mo ago
Business & Productivity ¶ TEXT

Meeting Minutes & Action Items from Transcript

Distills a raw meeting transcript into decisions, action items with owners and dates, and open questions — flagging anything ambiguous instead of guessing.

✓ Claude Haiku 4.5Gemini 2.5 Flash
★ 95⧉ 810◌ 0LLM Atlas Team · 2mo ago
LLM Atlas

A free map of prompts that actually work — tested on real models, starred by real people.

© 2026 LLM AtlasPrompts default to CC0 unless stated.