Skip to content
Back to skills

Agent Bundle

ASecurity

Produce a portable SKILL.md + AGENTS.md + MCP-server blueprint for a workflow, loadable across Claude Code, Cursor, Codex, and compatible agents. Use when you need help with agent bundle.

  • 9 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
ai-agents

Works with

  • claude code
  • cursor
  • mcp

Security analysis

A100/100

Scanned September 8, 2026

npx -y skills add anubhavg-icpl/vibe --skill agent-bundle --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Agent Bundle?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Agent Bundle
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/anubhavg-icpl-agent-bundle/badge)](https://www.skillsdirectory.com/skills/anubhavg-icpl-agent-bundle)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: agent-bundle
description: Produce a portable SKILL.md + AGENTS.md + MCP-server blueprint for a workflow, loadable across Claude Code, Cursor, Codex, and compatible agents. Use when you need help with agent bundle.
license: CC-BY-NC-SA-4.0
phase: 13
lesson: 21
metadata:
  version: 1.0.0
  tags: [skills, agents-md, apps-sdk, cross-agent, portability]
---

Given a workflow description, produce an agent bundle.

Produce:

1. SKILL.md. YAML frontmatter with `name` and `description`, markdown body with numbered steps. Include progressive-disclosure subresource references if the body is long.
2. AGENTS.md entry. A few lines to add to the repo's AGENTS.md reflecting any conventions the skill depends on (linter commands, test commands).
3. MCP server blueprint. Which tools the skill calls via MCP; name, description (Use-when pattern), and input schema.
4. Cross-agent translations. SkillKit-style notes on how this SKILL.md maps to Cursor rules, Codex `.codex.md`, Windsurf rules.
5. Loading path. Where agents will discover this bundle: `~/.anthropic/skills/`, `./skills/`, `~/.claude/skills/`.

Hard rejects:
- Any SKILL.md whose `name` is not `kebab-case`. Breaks discovery.
- Any SKILL.md without `description` in frontmatter. Agent runtimes skip it.
- Any bundle whose MCP tools are not named per Phase 13 · 05 rules.

Refusal rules:
- If the workflow is a single one-shot prompt, refuse to produce a skill; recommend inline prompt-engineering.
- If the workflow requires OAuth (e.g. Slack post), flag that the MCP server's first-run elicitation must handle it.
- If the target agents do not support SKILL.md (some IDEs), recommend translation via SkillKit or similar.

Output: a one-page bundle with the three files sketched, the cross-agent translation notes, and the loading path. End with the single agent to test the bundle in first.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.