How do I install the Write Vendor Config skill?

Run npx skillmds@latest add cyberuni/write-vendor-config in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Write Vendor Config safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Write Vendor Config?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Write Vendor Config free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Write Vendor Config?

cyberuni (@cyberuni) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Write Vendor Config by cyberuni — Agent Skill

Name: Write Vendor Config
Author: cyberuni

Write Vendor Config

Use this skill to write per-harness agent config files (CLAUDE.md, Cursor, Codex, Copilot) directly from AGENTS.md by hand, without the universal-plugin CLI. Use when wiring vendor config manually, when npx is unavailable or declined, or when a user asks to set up the harness config files directly rather than via the tool.

cyberuni 53129a7 2 files · 3.5 KB Updated Sep 17, 2026

File contents

write-vendor-config — write per-harness agent config directly

Write each targeted agent harness's configuration file from AGENTS.md by hand, without the universal-plugin CLI. This is the manual fallback the init skill routes to when the user declines running npx, and a standalone entry when someone wants the by-hand path.

The preferred wiring is init's CLI-offload step (universal-plugin sync) — if the user is fine running the CLI, that path is better and this skill defers to it. This skill exists for the no-CLI case.

1. Detect the harnesses in play

Write config only for the harnesses the project actually targets — do not blindly write all four. Treat any of these as a signal that a harness is targeted:

a vendor directory is present (.cursor/, .codex/, .github/),
a vendor config file already exists, or
the user names the harnesses to configure.

If no harness can be detected and the user named none, ask which harnesses to configure rather than writing files blindly.

2. Write each vendor's file from AGENTS.md

For each targeted harness, write its config file in that vendor's own shape, grounded in AGENTS.md rather than invented:

Harness	File
Claude Code	`CLAUDE.md`
Cursor	prefer `.cursor/rules/`; use a legacy `.cursorrules` only when the repo already uses it
Codex	the Codex agent-config file the repo uses
GitHub Copilot	`.github/copilot-instructions.md`

Derive each file's content from AGENTS.md; do not fabricate rules the source does not contain.

3. Stay idempotent

If a targeted vendor file already matches what you would write, leave it unchanged.
If it substantively differs, ask the user before overwriting it — never overwrite silently.

4. Never shell out to the CLI

This is the no-CLI path. Write the files directly with your file tools. Do not invoke npx universal-plugin or any universal-plugin subcommand — offloading to the CLI is init's step, not this skill's.

Report

State which harnesses were detected or named, which vendor files were written, updated, left unchanged, or skipped pending confirmation, and confirm no CLI was invoked.

---name: write-vendor-configdescription: Use this skill to write per-harness agent config files (CLAUDE.md, Cursor, Codex, Copilot) directly from AGENTS.md by hand, without the universal-plugin CLI. Use when wiring vendor config manually, when npx is unavailable or declined, or when a user asks to set up the harness config files directly rather than via the tool.---# write-vendor-config — write per-harness agent config directlyWrite each targeted agent harness's configuration file from `AGENTS.md` **by hand**, without the`universal-plugin` CLI. This is the manual fallback the `init` skill routes to when the user declines running`npx`, and a standalone entry when someone wants the by-hand path.The preferred wiring is `init`'s CLI-offload step (`universal-plugin sync`) — if the user is fine running theCLI, that path is better and this skill defers to it. This skill exists for the no-CLI case.## 1. Detect the harnesses in playWrite config only for the harnesses the project actually targets — do not blindly write all four. Treat any ofthese as a signal that a harness is targeted:- a vendor directory is present (`.cursor/`, `.codex/`, `.github/`),- a vendor config file already exists, or- the user names the harnesses to configure.If **no harness can be detected and the user named none**, ask which harnesses to configure rather than writingfiles blindly.## 2. Write each vendor's file from AGENTS.mdFor each targeted harness, write its config file in that vendor's own shape, grounded in `AGENTS.md` rather thaninvented:| Harness | File ||---|---|| Claude Code | `CLAUDE.md` || Cursor | prefer `.cursor/rules/`; use a legacy `.cursorrules` only when the repo already uses it || Codex | the Codex agent-config file the repo uses || GitHub Copilot | `.github/copilot-instructions.md` |Derive each file's content from `AGENTS.md`; do not fabricate rules the source does not contain.## 3. Stay idempotent- If a targeted vendor file **already matches** what you would write, leave it unchanged.- If it **substantively differs**, ask the user before overwriting it — never overwrite silently.## 4. Never shell out to the CLIThis is the no-CLI path. Write the files directly with your file tools. Do **not** invoke `npx universal-plugin`or any `universal-plugin` subcommand — offloading to the CLI is `init`'s step, not this skill's.## ReportState which harnesses were detected or named, which vendor files were written, updated, left unchanged, orskipped pending confirmation, and confirm no CLI was invoked.

Frequently asked questions