Install
$ agentstack add skill-claude-world-claude-agent-setup ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ● Network access Used
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Setup — First-Time Configuration
Welcome new users and configure claude-agent for their environment.
Steps
1. Welcome & Detect Environment
# Detect OS and shell
uname -s # Darwin / Linux
echo $SHELL
node --version
Present:
Welcome to Claude-Agent!
I'll help you set up in ~2 minutes. Here's what we'll do:
1. Learn about you (name, timezone, preferences)
2. Check which tools you have installed
3. Enable the skills that work on your system
4. Set up scheduled routines (optional)
5. Show you what's available
Detected: macOS / Linux | Node vX | zsh/bash
2. User Profile
Ask these questions and save to memory/user-profile.md:
- "What should I call you?"
- "What timezone are you in?" (auto-detect:
date +%Z) - "What language do you prefer?" (zh-TW / en / ja / other)
- "What's your role?" (developer / designer / PM / student / other)
- "What topics interest you most?" (for daily briefing & trend scout)
Update memory/user-profile.md with the answers.
3. Tool Inventory
Check which CLI tools are installed and categorize available skills:
# Check each tool — fast parallel checks
for cmd in gh himalaya gogcli memo remindctl things grizzly obsidian-cli \
spogo spotify_player sonos openhue eightctl camsnap ffmpeg \
whisper imsg wacli xurl op peekaboo jq rg nano-pdf gifgrep curl uvx; do
which $cmd 2>/dev/null && echo "FOUND: $cmd" || echo "MISSING: $cmd"
done
Present results grouped:
Tool Inventory:
READY (installed):
gh → /github-ops, /gh-issues
himalaya → /email
ffmpeg → /video-extract
curl → /weather, /quick-research (always available)
jq → /session-logs
rg → /session-logs
NOT INSTALLED (skills available after install):
gogcli → /google-workspace (brew install steipete/tap/gogcli)
grizzly → /bear-notes (go install github.com/tylerwince/grizzly/cmd/grizzly@latest)
spogo → /spotify (cargo install spogo)
openhue → /hue-lights (brew install openhue)
...
NO INSTALL NEEDED (always available):
/memory-manager /task-tracker /daily-briefing /brainstorm
/draft-message /quick-research /summarize /context-health
/weather /security-audit /skill-creator
4. MCP Server Status
Check which MCP servers are responding:
# Verify MCP servers in .mcp.json
cat .mcp.json | python3 -c "import json,sys; [print(k) for k in json.load(sys.stdin)['mcpServers']]"
Present:
MCP Servers:
trend-pulse (zero-auth) → /trend-scout, /content-creator
claude-101 (zero-auth) → 27 template tools
cf-browser (needs CF credentials) → /deep-read
notebooklm (needs Google login) → /podcast-maker, /content-creator with-image
Setup instructions for optional servers:
cf-browser: export CF_ACCOUNT_ID="..." CF_API_TOKEN="..." in ~/.zshrc
notebooklm: uvx notebooklm login (one-time browser sign-in)
5. Scheduled Routines
Ask if user wants recurring tasks:
Would you like to set up any recurring routines?
1. Memory consolidation — every few hours (recommended)
2. Daily briefing — every morning at a specific time
3. RSS feed check — periodic blog/news monitoring
4. Skip — set up later
Which ones? (1,2,3 or skip)
If user selects options, explain how to use /loop:
- Memory:
/loop 3h /memory-manager - Briefing: Explain the user can say "good morning" or set a reminder
- RSS:
/loop 1h /rss-monitor
6. Quick Tour
Show a categorized skill summary based on what's available:
Your Claude-Agent is ready! Here's what you can do:
CORE (always available):
"Good morning" → daily briefing with your tasks & news
"Remind me to X at Y" → task tracking with reminders
"Research X" → web research with summary
"Summarize " → condense any article
"Brainstorm about X" → structured ideation
"What do you remember?" → check your memory
CONTENT (MCP-powered):
"What's trending?" → 20 free trend sources
"Create a post about X" → patent-scored social content
"Make a podcast about X" → AI podcast via NotebookLM
INSTALLED ON YOUR SYSTEM:
[list only skills with installed CLI tools]
MESSAGING:
[list available messaging skills]
Type any of these naturally — I'll route to the right skill.
Run /context-health anytime to check system status.
7. Save Setup State
Write a marker to avoid re-prompting:
# In memory/user-profile.md, add:
## Setup
- Setup completed: YYYY-MM-DD
- OS: macOS/Linux
- Installed tools: [list]
- MCP servers: [list]
- Scheduled routines: [list or "none"]
Rules
- Be friendly and conversational — this is the first impression
- Don't overwhelm — show what's relevant to their system
- Skip tool checks if --skip-checks flag is set
- Minimal mode (--minimal): just ask name/timezone, skip tool inventory
- If user is migrating from OpenClaw, suggest
/migrate-openclawfirst - Save everything to memory immediately — don't wait
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: claude-world
- Source: claude-world/claude-agent
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.