CodingLlama
Llama Prompts for Coding
Llama 3.3 70B is a capable coding model that can be run locally or via cloud inference — making it the go-to choice for development teams that need AI coding assistance without sending code to third-party APIs. It's particularly popular for on-premises coding tools.
8 copy-paste ready prompts — optimised for Llama.
8 Llama Coding Prompts
Debug an Error
Language: [language + version]. Framework: [framework if applicable]. Error message: [paste exact error]. Relevant code: [paste code]. What I've tried: [attempts]. Please: 1. Identify the root cause. 2. Explain why this error occurs. 3. Provide a corrected version with inline comments. 4. Note any edge cases to handle.
Write a Function
Write a [language] function that [does X]. Requirements: [list specific requirements]. Input: [describe input type and format]. Output: [describe output type and format]. Include: docstring/JSDoc, input validation, error handling for [specific errors], and 3 unit tests covering normal case, edge case, and error case.
Code Review
Review this [language] code for: 1. Bugs or logic errors. 2. Security vulnerabilities (especially [SQL injection/XSS/SSRF etc. if relevant]). 3. Performance bottlenecks. 4. Readability improvements. 5. Missing error handling. Format: numbered list of issues with line references and concrete fix suggestions. [Paste code]
Refactor Code
Refactor this [language] code to improve [readability/performance/maintainability]. Apply [specific patterns: DRY/SOLID/etc.]. Do not change the external API or behaviour. Add comments to explain complex logic. Show the refactored version and list every change made with a one-sentence reason for each. [Paste code]
API Endpoint
Write a [framework] API endpoint for [method: GET/POST/etc.] at route [/path]. Purpose: [what it does]. Auth: [auth method]. Request body/params: [describe]. Response: [describe success + error format]. Include: input validation, error responses (400/401/404/500), and OpenAPI/Swagger comments.
Database Schema
Design a [database type: PostgreSQL/MySQL/MongoDB] schema for [application type]. Entities needed: [list entities]. Relationships: [describe]. Include: table definitions with field names, types, and constraints; foreign keys; indexes for performance; and a brief explanation of design decisions.
Write Unit Tests
Write unit tests for this [language] function using [testing framework]. Test: normal happy path, at least 2 edge cases, and all error conditions. Mock external dependencies. Each test should have a clear name describing what it verifies. [Paste the function]
Explain Code
Explain this [language] code as if I'm a developer who knows [beginner/intermediate] [language] but hasn't seen this pattern before. Walk through: 1. What it does overall. 2. Each section step by step. 3. Any non-obvious patterns or techniques used. 4. Potential gotchas. [Paste code]
Tips for using Llama for Coding
- Use the correct prompt template for your Llama variant — [INST]/[/INST] for older models, <|user|>/<|assistant|> for Llama 3
- For on-premises coding use, Llama handles code generation, review, and documentation reliably
- Llama 3.3 70B via Ollama is the most popular local coding AI setup — instructions work best in the Modelfile system prompt
Get a prompt tailored to your exact coding task
PromptIt asks a few questions and builds a Llama-optimised prompt for your specific use case — free.
✦ Try PromptIt Free