How do I install the Startup Context skill?

Run npx skillmds@latest add tomevault-io/startup-context 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 Startup Context 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 Startup Context?

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 Startup Context free to use?

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

Who published Startup Context?

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

Startup Context by tomevault-io — Agent Skill

Name: Startup Context
Author: tomevault-io

Startup Context

Validate and diagnose startup context files — CLAUDE.md, AGENTS.md, rules, and .cursorrules — that inject instructions into the model context at session start across Claude Code, Cursor, and Codex. Use when this capability is needed.

tomevault-io 793318e 2 files · 2.7 KB Updated Sep 17, 2026

File contents

Use this skill when working with files that inject instructions at startup — validating format, diagnosing why instructions aren't being read, or understanding load order and precedence per vendor.

What startup context is: Files that are automatically loaded into the model's context before any user prompt. Each vendor uses different file names and locations. All serve the same purpose: priming the model with project-specific or harness-level instructions.

Vendor mapping:

Vendor	Primary file	Rules format	Legacy
Claude Code	CLAUDE.md	.claude/rules/*.md	—
Cursor	.cursor/rules/*.mdc	.cursor/rules/*.mdc	.cursorrules (root)
Codex	AGENTS.md	NOT SUPPORTED in plugins	—

Load order (Claude Code): User CLAUDE.md (~/.claude/CLAUDE.md) → Project CLAUDE.md → Plugin CLAUDE.md (via @-import). Rules in .claude/rules/ are loaded after CLAUDE.md.

The @AGENTS.md workaround (Claude Code): Claude Code does not natively read AGENTS.md. When ynh exports for Claude, it writes a CLAUDE.md containing only @AGENTS.md — using Claude's @-import syntax to pull in the cross-vendor instructions file without duplicating content.

Diagnostic checklist:

Is the file in the right location for the vendor?
Claude Code: does the plugin CLAUDE.md use @-import correctly?
Cursor: is the rules file .mdc (not .md)? Does it have required frontmatter (description, globs, alwaysApply)?
Codex: rules are not supported in plugin format — use AGENTS.md only
Check file encoding and line endings — bare CR can cause silent failures

Source: eyelock/assistants — distributed by TomeVault.

---name: startup-contextdescription: Validate and diagnose startup context files — CLAUDE.md, AGENTS.md, rules, and .cursorrules — that inject instructions into the model context at session start across Claude Code, Cursor, and Codex. Use when this capability is needed.---Use this skill when working with files that inject instructions at startup — validating format, diagnosing why instructions aren't being read, or understanding load order and precedence per vendor.**What startup context is:** Files that are automatically loaded into the model's context before any user prompt. Each vendor uses different file names and locations. All serve the same purpose: priming the model with project-specific or harness-level instructions.**Vendor mapping:**| Vendor | Primary file | Rules format | Legacy ||--------|-------------|--------------|--------|| Claude Code | CLAUDE.md | .claude/rules/*.md | — || Cursor | .cursor/rules/*.mdc | .cursor/rules/*.mdc | .cursorrules (root) || Codex | AGENTS.md | NOT SUPPORTED in plugins | — |**Load order (Claude Code):** User CLAUDE.md (~/.claude/CLAUDE.md) → Project CLAUDE.md → Plugin CLAUDE.md (via @-import). Rules in .claude/rules/ are loaded after CLAUDE.md.**The @AGENTS.md workaround (Claude Code):** Claude Code does not natively read AGENTS.md. When ynh exports for Claude, it writes a CLAUDE.md containing only `@AGENTS.md` — using Claude's @-import syntax to pull in the cross-vendor instructions file without duplicating content.**Diagnostic checklist:**1. Is the file in the right location for the vendor?2. Claude Code: does the plugin CLAUDE.md use @-import correctly?3. Cursor: is the rules file .mdc (not .md)? Does it have required frontmatter (description, globs, alwaysApply)?4. Codex: rules are not supported in plugin format — use AGENTS.md only5. Check file encoding and line endings — bare CR can cause silent failures---> Source: [eyelock/assistants](https://github.com/eyelock/assistants) — distributed by [TomeVault](https://tomevault.io).

Frequently asked questions