Skip to content
Agentmart
Submit a tool

code-writer

Build, modify, and verify local code deliverables in common languages, especially Python, JavaScript, TypeScript, HTML/CSS, SQL, and shell scripts.

Skill by TheRefreshCNFT

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/therefreshcnft-code-writer

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Build, modify, and verify local code deliverables in common languages, especially Python, JavaScript, TypeScript, HTML/CSS, SQL, and shell scripts. Use when the user asks to write a script, CLI, module, web app, API client, parser, test, automation, or other executable/source-code artifact. Do not use as a substitute for SocratiCode/Axon when the task is primarily codebase discovery, semantic search, call graph, dependency impact, or dead-code analysis.

Use cases: frontend, backend, testing, writing

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/code-writer

What's inside

Skill name
code-writer
Description
Build, modify, and verify local code deliverables in common languages, especially Python, JavaScript, TypeScript, HTML/CSS, SQL, and shell scripts. Use when the user asks to write a script, CLI, module, web app, API client, parser, test, automation, or other executable/source-code artifact. Do not use as a substitute for SocratiCode/Axon when the task is primarily codebase discovery, semantic search, call graph, dependency impact, or dead-code analysis.
License
MIT
Instructions
56 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Code Writer

Use this skill when the primary deliverable is source code or an executable
local artifact.

## Routing

- Fresh code, scripts, CLIs, modules, utilities, tests, parsers, small web apps,
  API clients, SQL, shell scripts: use this skill.
- Existing-codebase discovery, architecture mapping, "where does this live",
  semantic search, dependency context: use SocratiCode first.
- Structural graph, call graph, blast radius, circular dependencies, dead code:
  use Axon.
- UI, dashboard, layout, component states, visual hierarchy, accessibility:
  pair with UI/UX Pro Max. UI/UX owns design direction; Code Writer owns the
  concrete implementation.
- Web scraping/crawling/browser extraction: pair with Scrapling. Scrapling owns
  source acquisition; Code Writer owns parsing, transformation, and packaging.

## Core Workflow

1. Identify the runnable unit: script, CLI, module, webpage, test, SQL, or config.
2. Choose the smallest conventional structure for the language and task.
3. Write complete files, not patches or placeholders.
4. Add error handling for expected bad inputs and missing files.
5. Keep dependencies minimal. Prefer the standard library unless a package is
   clearly justified and installable in the workspace.
6. Include a simple validation path: unit test, smoke command, syntax check,
   sample input/output, or deterministic assertion.
7. If workspace exec is available, request a simple command in `COMMANDS` and
   do not claim it ran until the harness records the command run.

## Language Defaults

- Python: prefer standard-library CLIs with `argparse`, clear functions, and
  `if __name__ == "__main__": main()`. Use `pathlib`, structured exceptions,
  and tests or sample commands.
- JavaScript/TypeScript: prefer small ES modules or browser-safe scripts.
  Avoid framework scaffolds unless the user asks for a framework.
- HTML/CSS/JS: produce a complete runnable page. Keep links/assets
  workspace-relative and emit every referenced local file.
- SQL: write explicit schema/query files with comments only where they clarify
  intent. Do not claim execution unless a database command actually ran.
- Shell: keep scripts POSIX-ish when possible, quote variables, use `set -euo
  pipefail` for Bash/Zsh scripts, and avoid destructive commands.

## Verification Rules

- Syntax checks are not enough for user-facing behavior; include a behavioral
  smoke check when possible.
- For generated files, verify paths and referenced assets exist.
- For parsing/transforms, include a sample input and expected output or a small
  assertion.
- For UI code, pair with screenshots when the harness/browser path supports it.

From TheRefreshCNFT/gemma-forge, MIT.

More from TheRefreshCNFT/gemma-forge

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
May 24, 2026 (5 months ago)
Last release
May 25, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from TheRefreshCNFT

Cite this page

code-writer skill. Agent Mart. https://agentmart.dev/skills/therefreshcnft-code-writer (last updated 2026-10-24)