Qwen Coder
Delegate coding, analysis, and documentation runs to the Qwen Coder CLI with the right model and flags from your main agent workflow.
Overview
Qwen Coder is an agent skill for the Build phase that documents CLI delegation to Qwen2.5-Coder and QwQ with prompts, sessions, and model-specific flags.
Install
npx skills add https://github.com/giuseppe-trisciuoglio/developer-kit --skill qwen-coderWhat is this skill?
- Non-interactive delegation via `qwen -p "<prompt>"` for agent handoffs
- Session continue/resume with `-c` and `-r` for multi-step analysis
- Model split: Qwen2.5-Coder for fast production code and docs; QwQ for deeper reasoning tasks
- Approval modes (`auto_edit`, `plan`) paired with model choice in examples
- Structured JSON output patterns for automation and tooling integration
Adoption & trust: 707 installs on skills.sh; 271 GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You want your main agent to call Qwen for codegen or analysis but keep forgetting session flags, models, and non-interactive `-p` usage.
Who is it for?
Indie builders stacking multiple LLM CLIs who need a stable delegation cheat sheet during active development.
Skip if: Teams that do not install or allow a local `qwen` binary, or builders who only use in-chat codegen with no external CLI.
When should I use this skill?
You need to run Qwen Coder from scripts or agent skills with the correct model, approval mode, and session flags.
What do I get? / Deliverables
You get copy-paste CLI patterns for delegated runs, session continuity, and model selection aligned to task type.
- Delegated CLI runs (code, docs, or JSON analysis)
- Continued or resumed Qwen sessions by ID
- Model-matched prompts for production code versus reasoning
Recommended Skills
Journey fit
The skill documents how to invoke an external coding model CLI during implementation—not research or launch work. It sits on agent-tooling because it wires Prism-style delegation to Qwen sessions, models, and approval modes rather than shipping product UI.
How it compares
Use as a command reference for Qwen delegation, not as a Vue/React coding style guide or an MCP server integration.
Common Questions / FAQ
Who is qwen-coder for?
Solo and indie builders using agent harnesses who delegate subtasks to the Qwen Coder CLI alongside Claude Code, Cursor, or Codex.
When should I use qwen-coder?
During Build when generating endpoints, docs, or JSON analysis via `qwen -p`, when resuming long sessions with `-c`/`-r`, or when choosing Qwen2.5-Coder versus QwQ for speed versus reasoning.
Is qwen-coder safe to install?
It is procedural documentation only; review the Security Audits panel on this Prism page before trusting the parent developer-kit repo and any CLI you install.
SKILL.md
READMESKILL.md - Qwen Coder
# Qwen Coder CLI Command Reference Quick reference for the `qwen-coder` skill delegation workflow. ## Core Usage ```bash # Interactive mode qwen # Non-interactive single prompt (recommended for delegation) qwen -p "<english-prompt>" ``` ## Prompting and Sessions ```bash # Non-interactive prompt qwen -p "Analyze this codebase architecture" # Continue previous session by ID qwen -c <session-id> -p "Continue analysis and focus on auth module" # Resume previous session by ID qwen -r <session-id> -p "Continue analysis and focus on auth module" ``` ## Model Selection Choose the right model based on task requirements. Qwen2.5-Coder and QwQ are designed for complementary use cases: ### Qwen2.5-Coder — Fast, Production-Qode Focus ```bash # Code generation with production quality # Best for: boilerplate, CRUD endpoints, standard patterns qwen -p "Generate a production-ready REST API endpoint for CRUD operations on items. Include input validation, error handling, and unit tests. Use Express.js framework." -m qwen2.5-coder --approval-mode auto_edit # Documentation generation # Best for: JSDoc, README sections, API docs qwen -p "Generate comprehensive documentation for the UserService class. Include: class purpose, public methods with parameters, usage examples, and error handling patterns. Format as markdown." -m qwen2.5-coder --approval-mode plan # Structured JSON output for automation # Best for: data extraction, code analysis, tooling integration qwen -p "Analyze this codebase and return the top 5 refactoring opportunities as a JSON array. Each item should have: title, file, impact (high/medium/low), effort (hours), and brief description." -m qwen2.5-coder --output-format json ``` ### QwQ — Deep Reasoning for Complex Tasks ```bash # Deep security analysis # Best for: thorough security reviews, vulnerability assessment qwen -p "Analyze the authentication module for security vulnerabilities. Report only high-confidence issues with severity, file paths, and remediation steps." -m qwq --approval-mode plan # Architectural design reasoning # Best for: system design, migration planning, complex refactoring qwen -p "Analyze the current system architecture and propose a detailed migration strategy to microservices. Include component boundaries, communication patterns, data ownership, and estimated risks for each migration step." -m qwq --approval-mode plan # Complex multi-file refactoring # Best for: cross-module changes, architectural refactors, deep reasoning qwen -p "Refactor the payment module for better error handling. Consider transaction boundaries, retry logic, and idempotency. Propose a detailed plan before implementing." -m qwq --approval-mode plan ``` ### Performance and Model Tradeoffs | Model | Speed | Best For | |-------|-------|----------| | **qwen2.5-coder** | Fast | Code generation, documentation, automation, quick iterations | | **qwq** | Slower | Deep analysis, architectural reasoning, complex security reviews | **Selection guidance**: - Default to `qwen2.5-coder` for speed and production-quality code generation. - Use `qwq` when the task requires deep reasoning, architectural analysis, or thorough security reviews. - Use `--output-format json` for reliable machine-readable results in pipelines. ## Approval Modes ```bash # Default approval (recommended baseline) qwen -p "<english-prompt>" --approval-mode default # Auto-approve edits qwen -p "<english-prompt>" --approval-mode auto_edit # Read-only planning mode (preferred for analysis) qwen -p "<english-prompt>" --approval-mode plan # YOLO mode (only with explicit user consent) qwen -p "<english-prompt>" --approval-mode yolo ``` ## Output Control ```bash # Plain text (default) qwen -p "<english-prompt>" # JSON output for automation qwen -p "<english-prompt>" --output-format json # Streaming JSON output qwen -p "<english-prompt>" --output-format stream-json ``` ## Safe Delegation Patterns ```bash # Security review in read-only mode qwen -p "Analy