feat: add installable looper agent skill

[mrcfps]

Summary

Add an installable agent skill for Looper so users can install it with npx skills add and let agents configure, run, and troubleshoot looper / looperd workflows.

Motivation

Users should be able to make Looper discoverable and usable by agent environments without manually copying project instructions. A dedicated skill can encode Looper-specific setup, configuration, daemon lifecycle, command usage, and troubleshooting guidance.

Proposed scope

Create a Looper skill with a structure similar to:

skills/looper/
  SKILL.md
  references/
    config.md
    cli.md
    daemon.md
  scripts/
    check.sh

The skill should cover:

• Installing/adding the skill via npx skills add
• Using Looper from agents
• Supported commands:
  • go run ./cmd/looperd
  • go run ./cmd/looper <args>
  • go build ./...
  • go vet ./...
  • go test ./...
• Default config path: ~/.looper/config.json
• Config precedence: defaults → config file → env → CLI flags
• Runtime artifacts under ~/.looper/
• Startup validation and writable runtime path requirements
• Tool auto-detection for git, gh, and osascript
• osascript startup requirement when notifications are enabled
• Safety guidance for agents, including not overwriting user config without confirmation

Skill discovery requirements

SKILL.md should include concise frontmatter:

---
name: looper
description: Looper CLI/daemon — use when configuring, starting, checking, or operating looper, looperd, or ~/.looper config/runtime paths. Use for setup, status, logs, config, and troubleshooting.
---

The description should include trigger language such as:

• configure looper
• start looperd
• check looper status
• debug looper config
• fix writable runtime paths
• use looper via agents

Acceptance criteria

• A looper agent skill exists in the repository.
• The skill can be installed by users with npx skills add from the repository or relevant skill path.
• The skill includes SKILL.md with valid name and description frontmatter.
• The skill documents Looper setup, configuration, daemon lifecycle, CLI usage, and troubleshooting.
• The skill includes safety notes for config changes and daemon startup side effects.
• Any helper script is safe, deterministic, and documented.
• Documentation explains how users install and verify the skill.

References

• https://github.com/vercel-labs/add-skill/blob/main/README.md
• https://developers.openai.com/codex/skills
• https://agentskills.io/specification

[mrcfps]

Added in skills/looper.