neoagent 1.0.0

package/server/routes/store.js

@@ -0,0 +1,1192 @@
+const express = require('express');
+const router = express.Router();
+const path = require('path');
+const { requireAuth } = require('../middleware/auth');
+
+router.use(requireAuth);
+
+const SKILLS_DIR = path.join(__dirname, '../../agent-data/skills');
+
+// ── Skill catalog ─────────────────────────────────────────────────────────────
+// Each entry: id (becomes filename <id>.md), name, description, category, icon, content
+const CATALOG = [
+
+  // ── SYSTEM ──────────────────────────────────────────────────────────────────
+  {
+    id: 'disk-usage',
+    name: 'Disk Usage',
+    description: 'Show disk space usage for all mounted filesystems.',
+    category: 'system',
+    icon: '💾',
+    content: `---
+name: disk-usage
+description: Show disk space usage for all mounted filesystems
+category: system
+icon: 💾
+enabled: true
+---
+
+Run \`df -h\` to show disk usage. If the user asks about a specific path, run \`du -sh <path>\`. Present the output in a readable table, highlight any filesystems above 80% usage.`
+  },
+
+  {
+    id: 'system-stats',
+    name: 'System Stats',
+    description: 'Report CPU, RAM, load average and uptime.',
+    category: 'system',
+    icon: '📊',
+    content: `---
+name: system-stats
+description: Report CPU, RAM, load average and uptime
+category: system
+icon: 📊
+enabled: true
+---
+
+Run these commands and summarise the results:
+- \`uptime\` — load average and uptime
+- \`free -m\` (Linux) or \`vm_stat\` (macOS) — memory usage
+- \`nproc\` or \`sysctl -n hw.logicalcpu\` — CPU count
+
+Combine into a short dashboard-style summary.`
+  },
+
+  {
+    id: 'process-monitor',
+    name: 'Process Monitor',
+    description: 'List the top 15 processes by CPU or memory usage.',
+    category: 'system',
+    icon: '⚙️',
+    content: `---
+name: process-monitor
+description: List the top 15 processes by CPU or memory usage
+category: system
+icon: ⚙️
+enabled: true
+---
+
+Run \`ps aux --sort=-%cpu | head -16\` on Linux or \`ps aux -r | head -16\` on macOS to show top CPU processes. If the user asks for memory, use \`ps aux --sort=-%mem | head -16\`. Format as a readable table with PID, command, CPU%, MEM%.`
+  },
+
+  {
+    id: 'tail-log',
+    name: 'Tail Log File',
+    description: 'Show the last N lines of any log file, with optional filtering.',
+    category: 'system',
+    icon: '📋',
+    content: `---
+name: tail-log
+description: Show the last N lines of any log file, with optional filtering
+category: system
+icon: 📋
+enabled: true
+---
+
+Run \`tail -n <lines> <file>\` to show recent log lines. Default to 50 lines if not specified. If the user provides a filter keyword, pipe through \`grep <keyword>\`. For common logs, suggest: /var/log/syslog, /var/log/nginx/error.log, ./logs/app.log etc.`
+  },
+
+  {
+    id: 'find-large-files',
+    name: 'Find Large Files',
+    description: 'Find the largest files in a directory tree.',
+    category: 'system',
+    icon: '🔍',
+    content: `---
+name: find-large-files
+description: Find the largest files in a directory tree
+category: system
+icon: 🔍
+enabled: true
+---
+
+Run \`find <dir> -type f -exec du -sh {} + 2>/dev/null | sort -rh | head -20\` to list the 20 largest files. Default to current directory if none provided. Present as a ranked list with human-readable sizes.`
+  },
+
+  // ── NETWORK ─────────────────────────────────────────────────────────────────
+  {
+    id: 'ping-host',
+    name: 'Ping Host',
+    description: 'Ping a hostname or IP and report latency and packet loss.',
+    category: 'network',
+    icon: '📡',
+    content: `---
+name: ping-host
+description: Ping a hostname or IP and report latency and packet loss
+category: network
+icon: 📡
+enabled: true
+---
+
+Run \`ping -c 5 <host>\` to send 5 ICMP packets. Report average RTT, packet loss %, and whether the host is reachable. If the host is unreachable, suggest checking DNS or firewall.`
+  },
+
+  {
+    id: 'ip-info',
+    name: 'IP Info',
+    description: 'Get your public IP address and geolocation details.',
+    category: 'network',
+    icon: '🌐',
+    content: `---
+name: ip-info
+description: Get your public IP address and geolocation details
+category: network
+icon: 🌐
+enabled: true
+---
+
+Make a GET request to \`https://ipinfo.io/json\` (no auth needed for basic info). Display the IP, city, region, country, org (ISP), and timezone in a clean summary. If the user asks about a specific IP, use \`https://ipinfo.io/<ip>/json\`.`
+  },
+
+  {
+    id: 'ssl-check',
+    name: 'SSL Certificate Check',
+    description: 'Check the SSL certificate expiry date for any domain.',
+    category: 'network',
+    icon: '🔒',
+    content: `---
+name: ssl-check
+description: Check the SSL certificate expiry date for any domain
+category: network
+icon: 🔒
+enabled: true
+---
+
+Run \`echo | openssl s_client -connect <domain>:443 -servername <domain> 2>/dev/null | openssl x509 -noout -dates\` to get cert validity dates. Calculate how many days until expiry. Warn if < 30 days, alert if < 7 days, or if already expired.`
+  },
+
+  {
+    id: 'port-check',
+    name: 'Port Check',
+    description: 'Test whether a specific TCP port is open on a host.',
+    category: 'network',
+    icon: '🔌',
+    content: `---
+name: port-check
+description: Test whether a specific TCP port is open on a host
+category: network
+icon: 🔌
+enabled: true
+---
+
+Run \`nc -zv -w3 <host> <port> 2>&1\` or \`curl -s --connect-timeout 3 telnet://<host>:<port>\` to test connectivity. Report clearly: open or closed/filtered, and response time if measurable. Common ports to suggest: 80 (HTTP), 443 (HTTPS), 22 (SSH), 3306 (MySQL), 5432 (Postgres), 6379 (Redis).`
+  },
+
+  {
+    id: 'dns-lookup',
+    name: 'DNS Lookup',
+    description: 'Perform DNS lookups (A, MX, TXT, CNAME records) for any domain.',
+    category: 'network',
+    icon: '🗺️',
+    content: `---
+name: dns-lookup
+description: Perform DNS lookups (A, MX, TXT, CNAME records) for any domain
+category: network
+icon: 🗺️
+enabled: true
+---
+
+Use \`dig <domain> <type>\` or \`nslookup\` to query DNS records. Default to A records. If the user says "all records", run dig for A, MX, TXT, CNAME in one go. Present results cleanly without raw dig headers.`
+  },
+
+  // ── INFO ────────────────────────────────────────────────────────────────────
+  {
+    id: 'weather',
+    name: 'Weather',
+    description: 'Get current weather and a 3-day forecast for any city.',
+    category: 'info',
+    icon: '🌤️',
+    content: `---
+name: weather
+description: Get current weather and a 3-day forecast for any city
+category: info
+icon: 🌤️
+enabled: true
+---
+
+Make a GET request to \`https://wttr.in/<city>?format=j1\` (returns JSON). Parse:
+- current_condition[0]: temp_C, weatherDesc, humidity, windspeedKmph, feels_like
+- weather[0..2]: date, maxtempC, mintempC, hourly[4].weatherDesc (midday)
+
+Present as a clean weather card: current conditions + 3-day forecast with icons. URL-encode city names with spaces.`
+  },
+
+  {
+    id: 'crypto-price',
+    name: 'Crypto Price',
+    description: 'Look up live cryptocurrency prices from CoinGecko.',
+    category: 'info',
+    icon: '₿',
+    content: `---
+name: crypto-price
+description: Look up live cryptocurrency prices from CoinGecko
+category: info
+icon: ₿
+enabled: true
+---
+
+Use the CoinGecko free API (no key needed):
+- Single coin: \`GET https://api.coingecko.com/api/v3/simple/price?ids=<id>&vs_currencies=usd,eur&include_24hr_change=true\`
+- Common IDs: bitcoin, ethereum, solana, cardano, dogecoin, ripple, polkadot, chainlink, litecoin, avalanche-2
+
+Show price in USD and EUR, 24h change %, and a trend arrow ↑↓. If the user uses a ticker (BTC), map it to the CoinGecko ID first.`
+  },
+
+  {
+    id: 'exchange-rate',
+    name: 'Exchange Rate',
+    description: 'Get live currency exchange rates between any two currencies.',
+    category: 'info',
+    icon: '💱',
+    content: `---
+name: exchange-rate
+description: Get live currency exchange rates between any two currencies
+category: info
+icon: 💱
+enabled: true
+---
+
+Use the free Open Exchange Rates API:
+\`GET https://open.er-api.com/v6/latest/<base_currency>\`
+
+Example: \`https://open.er-api.com/v6/latest/USD\` returns rates for all currencies relative to USD.
+Show the requested conversion with the exact rate and the last updated time. If the user gives an amount (e.g. "200 EUR to GBP"), calculate and show the converted value.`
+  },
+
+  {
+    id: 'world-time',
+    name: 'World Time',
+    description: 'Show the current local time in major cities around the world.',
+    category: 'info',
+    icon: '🕐',
+    content: `---
+name: world-time
+description: Show the current local time in major cities around the world
+category: info
+icon: 🕐
+enabled: true
+---
+
+Run \`date\` for local time. Get world times via the API:
+\`GET https://worldtimeapi.org/api/timezone/<Region/City>\`
+
+Show a table of current times for: New York (America/New_York), London (Europe/London), Berlin (Europe/Berlin), Dubai (Asia/Dubai), Singapore (Asia/Singapore), Tokyo (Asia/Tokyo), Sydney (Australia/Sydney). Format as HH:MM timezone with day name.`
+  },
+
+  {
+    id: 'news-hackernews',
+    name: 'Hacker News Top Stories',
+    description: 'Fetch the top 10 stories from Hacker News right now.',
+    category: 'info',
+    icon: '📰',
+    content: `---
+name: news-hackernews
+description: Fetch the top 10 stories from Hacker News right now
+category: info
+icon: 📰
+enabled: true
+---
+
+1. GET \`https://hacker-news.firebaseio.com/v0/topstories.json\` → get array of IDs
+2. Take the first 10 IDs
+3. For each ID, GET \`https://hacker-news.firebaseio.com/v0/item/<id>.json\` → title, url, score, by, descendants
+Present as a numbered list: score points | title | by author (N comments). Link the title.`
+  },
+
+  // ── DEV ─────────────────────────────────────────────────────────────────────
+  {
+    id: 'git-summary',
+    name: 'Git Summary',
+    description: 'Show recent commits, current branch, and status for a git repo.',
+    category: 'dev',
+    icon: '🌿',
+    content: `---
+name: git-summary
+description: Show recent commits, current branch, and status for a git repo
+category: dev
+icon: 🌿
+enabled: true
+---
+
+Run in the user's specified directory (default: current working directory):
+1. \`git log --oneline -10\` — last 10 commits
+2. \`git status --short\` — dirty files
+3. \`git branch --show-current\` — current branch
+4. \`git remote -v\` — remotes
+
+Present as a structured git dashboard. Note any uncommitted changes or detached HEAD.`
+  },
+
+  {
+    id: 'docker-status',
+    name: 'Docker Status',
+    description: 'List all running and stopped Docker containers with their status.',
+    category: 'dev',
+    icon: '🐳',
+    content: `---
+name: docker-status
+description: List all running and stopped Docker containers with their status
+category: dev
+icon: 🐳
+enabled: true
+---
+
+Run \`docker ps -a --format "table {{.Names}}\\t{{.Image}}\\t{{.Status}}\\t{{.Ports}}"\` to list all containers. Also run \`docker images --format "table {{.Repository}}:{{.Tag}}\\t{{.Size}}"\` to show local images.
+Highlight running vs stopped vs exited. If Docker isn't installed/running, say so clearly.`
+  },
+
+  {
+    id: 'npm-outdated',
+    name: 'NPM Outdated Check',
+    description: 'Check for outdated npm packages in a Node.js project.',
+    category: 'dev',
+    icon: '📦',
+    content: `---
+name: npm-outdated
+description: Check for outdated npm packages in a Node.js project
+category: dev
+icon: 📦
+enabled: true
+---
+
+Run \`npm outdated --json\` in the project directory. Parse the JSON and display as a table:
+Package | Current | Wanted | Latest | Type (dep/devDep)
+
+Categorise: minor updates (wanted > current), major updates (latest >> current). Suggest running \`npm update\` for minor updates or manual upgrades for majors. Handle "everything up to date" gracefully.`
+  },
+
+  {
+    id: 'run-tests',
+    name: 'Run Tests',
+    description: 'Run the test suite for a project and summarise results.',
+    category: 'dev',
+    icon: '✅',
+    content: `---
+name: run-tests
+description: Run the test suite for a project and summarise results
+category: dev
+icon: ✅
+enabled: true
+---
+
+Check what test runner is configured (look at package.json scripts.test). Then run \`npm test\` (or \`yarn test\`, \`pytest\`, etc. as appropriate). Capture stdout/stderr.
+Summarise: total tests, passed, failed, skipped. If failures occur, show the first 3 failed test names and errors. Do NOT truncate error details — they're important.`
+  },
+
+  {
+    id: 'http-debug',
+    name: 'HTTP Debug',
+    description: 'Make a detailed HTTP request and inspect headers, status, timing.',
+    category: 'dev',
+    icon: '🔎',
+    content: `---
+name: http-debug
+description: Make a detailed HTTP request and inspect headers, status, timing
+category: dev
+icon: 🔎
+enabled: true
+---
+
+Use the http_request tool to make the request with full header capture. Also run \`curl -o /dev/null -s -w "\\n%{http_code} | %{time_total}s | %{size_download} bytes\\n" <url>\` for timing. Report:
+- Status code and meaning
+- Response time
+- Key headers (Content-Type, Cache-Control, X-Frame-Options, etc.)
+- Body preview (first 500 chars)
+- Any redirects`
+  },
+
+  // ── PRODUCTIVITY ─────────────────────────────────────────────────────────────
+  {
+    id: 'summarize-url',
+    name: 'Summarize URL',
+    description: 'Fetch a webpage and give a concise summary of its content.',
+    category: 'productivity',
+    icon: '📄',
+    content: `---
+name: summarize-url
+description: Fetch a webpage and give a concise summary of its content
+category: productivity
+icon: 📄
+enabled: true
+---
+
+Use http_request to GET the URL. Extract text content from the HTML (skip scripts, styles, nav). Write a structured summary:
+- **What it is**: 1 sentence
+- **Key points**: 3–5 bullet points
+- **Takeaway**: most important thing to know
+
+Keep total summary under 200 words. If the URL fails or returns non-HTML, say so.`
+  },
+
+  {
+    id: 'wikipedia',
+    name: 'Wikipedia Summary',
+    description: 'Get a Wikipedia article summary for any topic.',
+    category: 'productivity',
+    icon: '📚',
+    content: `---
+name: wikipedia
+description: Get a Wikipedia article summary for any topic
+category: productivity
+icon: 📚
+enabled: true
+---
+
+Use the Wikipedia REST API:
+\`GET https://en.wikipedia.org/api/rest_v1/page/summary/<title>\`
+
+URL-encode the title (replace spaces with underscores). The response contains \`extract\` (plain text summary) and \`content_urls.desktop.page\` (full article link).
+
+Show the summary with the article URL. If the topic resolves to a disambiguation page, show the top options. Support language override via \`https://<lang>.wikipedia.org/...\`.`
+  },
+
+  {
+    id: 'translate',
+    name: 'Translate Text',
+    description: 'Translate any text to a target language.',
+    category: 'productivity',
+    icon: '🌍',
+    content: `---
+name: translate
+description: Translate any text to a target language
+category: productivity
+icon: 🌍
+enabled: true
+---
+
+Use the LibreTranslate free API:
+\`POST https://libretranslate.com/translate\`
+Body: \`{"q": "<text>", "source": "auto", "target": "<lang_code>", "format": "text"}\`
+
+Common language codes: en, de, fr, es, it, pt, nl, ru, zh, ja, ko, ar.
+Show the translation clearly, note the detected source language. If the API fails, fall back to using your own translation ability with a note.`
+  },
+
+  {
+    id: 'quick-note',
+    name: 'Quick Note',
+    description: 'Save a timestamped note to a notes file on disk.',
+    category: 'productivity',
+    icon: '📝',
+    content: `---
+name: quick-note
+description: Save a timestamped note to a notes file on disk
+category: productivity
+icon: 📝
+enabled: true
+---
+
+Append the note to \`~/notes.md\` (or a user-specified file) with this format:
+\`\`\`
+## 2025-01-15 14:32
+<note content>
+\`\`\`
+Use \`echo\` or \`tee -a\` to append. Confirm the note was saved and show the file path. If the file doesn't exist, create it with a \`# Notes\` header first.`
+  },
+
+  {
+    id: 'pomodoro',
+    name: 'Pomodoro Timer',
+    description: 'Start a Pomodoro focus timer with a desktop notification at the end.',
+    category: 'productivity',
+    icon: '🍅',
+    content: `---
+name: pomodoro
+description: Start a Pomodoro focus timer with a desktop notification at the end
+category: productivity
+icon: 🍅
+enabled: true
+---
+
+Default: 25-minute work session followed by a 5-minute break. Run in background:
+\`\`\`bash
+(sleep 1500 && osascript -e 'display notification "Pomodoro complete! Take a break." with title "🍅 Pomodoro"' 2>/dev/null || notify-send "🍅 Pomodoro" "Complete! Take a break." 2>/dev/null || echo "POMODORO DONE") &
+\`\`\`
+Print the PID and end time so the user can track it. Support custom durations.`
+  },
+
+  // ── GOOGLE ──────────────────────────────────────────────────────────────────
+  {
+    id: 'gogcli',
+    name: 'Google Suite (gogcli)',
+    description: 'Control Gmail, Calendar, Drive, Tasks, Contacts, Sheets, Chat and more via the gog CLI.',
+    category: 'productivity',
+    icon: '🌐',
+    content: `---
+name: gogcli
+description: Control Gmail, Calendar, Drive, Tasks, Contacts, Sheets, Chat and more via the gog CLI
+category: productivity
+icon: 🌐
+trigger: gmail|calendar|drive|tasks|contacts|sheets|google
+enabled: true
+source: https://github.com/steipete/gogcli
+---
+
+## Overview
+\`gog\` (gogcli) is a fast, JSON-first CLI for Google Workspace services. Install via:
+\`\`\`
+brew install steipete/tap/gogcli
+\`\`\`
+Use \`--json\` for machine-readable output.
+
+## Credential Storage (NeoAgent)
+Store the Google account email in API_KEYS.json so you can always retrieve it:
+\`\`\`
+memory_write target=api_keys content={"gog_account": "you@gmail.com"}
+\`\`\`
+For non-interactive / headless runs, also store the keyring password:
+\`\`\`
+memory_write target=api_keys content={"gog_account": "you@gmail.com", "gog_keyring_password": "<your-keyring-password>"}
+\`\`\`
+Before every \`gog\` command, read back these keys:
+\`\`\`
+memory_read target=api_keys
+\`\`\`
+Then run gog with the stored values:
+\`\`\`bash
+export GOG_ACCOUNT="<gog_account from api_keys>"
+export GOG_KEYRING_BACKEND=file                       # use on-disk encrypted keyring (no Keychain prompts)
+export GOG_KEYRING_PASSWORD="<gog_keyring_password from api_keys>"  # if set
+gog ...
+\`\`\`
+
+## First-Time Setup (run once manually by the user)
+\`\`\`bash
+gog auth credentials ~/Downloads/client_secret_....json  # store OAuth client credentials
+gog auth keyring file                                     # switch to file backend (avoids Keychain prompts)
+gog auth add you@gmail.com                               # authorize account — opens browser
+gog auth status                                          # verify auth state
+gog auth list                                            # list stored accounts
+\`\`\`
+After setup, save the account email with memory_write (above) so the agent can use it automatically.
+
+## Gmail
+\`\`\`bash
+gog gmail search 'newer_than:7d is:unread' --max 10 --json
+gog gmail thread get <threadId>
+gog gmail send --to a@b.com --subject "Subject" --body "Body"
+gog gmail send --to a@b.com --subject "Subject" --body-html "<p>Hi</p>"
+gog gmail labels list
+gog gmail thread modify <threadId> --add STARRED --remove INBOX
+\`\`\`
+
+## Calendar
+\`\`\`bash
+gog calendar events primary --today --json
+gog calendar events primary --week
+gog calendar search "standup" --days 7 --json
+gog calendar create primary --summary "Meeting" --from 2026-01-15T10:00:00Z --to 2026-01-15T11:00:00Z --attendees "a@b.com"
+gog calendar freebusy --calendars primary --from 2026-01-15T00:00:00Z --to 2026-01-16T00:00:00Z
+gog calendar delete primary <eventId> --force
+\`\`\`
+
+## Drive
+\`\`\`bash
+gog drive ls --max 20 --json
+gog drive search "invoice" --max 20 --json
+gog drive upload ./file.pdf
+gog drive download <fileId> --out ./file.pdf
+gog drive share <fileId> --to user --email user@example.com --role reader
+\`\`\`
+
+## Tasks
+\`\`\`bash
+gog tasks lists --json
+gog tasks list <tasklistId> --json
+gog tasks add <tasklistId> --title "Task title" --due 2026-02-01
+gog tasks done <tasklistId> <taskId>
+\`\`\`
+
+## Contacts
+\`\`\`bash
+gog contacts search "John" --max 20 --json
+gog contacts create --given "Jane" --family "Doe" --email "jane@example.com"
+\`\`\`
+
+## Sheets
+\`\`\`bash
+gog sheets metadata <spreadsheetId>
+gog sheets get <spreadsheetId> 'Sheet1!A1:D10' --json
+gog sheets update <spreadsheetId> 'Sheet1!A1' 'val1|val2,val3|val4'
+gog sheets append <spreadsheetId> 'Sheet1!A:C' 'new|row|data'
+\`\`\`
+
+## Docs / Slides
+\`\`\`bash
+gog docs cat <docId>
+gog docs export <docId> --format pdf --out ./doc.pdf
+gog docs sed <docId> 's/old/new/g'
+gog slides export <presentationId> --format pptx --out ./deck.pptx
+\`\`\`
+
+## Tips
+- Pipe JSON output to \`jq\` for filtering: \`gog --json gmail search 'is:unread' | jq '.threads[].id'\`
+- Use \`GOG_ENABLE_COMMANDS=gmail,calendar\` to sandbox allowed commands
+- \`--plain\` outputs stable TSV for shell scripting
+- Calendar JSON includes \`startDayOfWeek\` and localized time fields, useful for scheduling logic`
+  },
+
+  // ── COMMUNITY ────────────────────────────────────────────────────────────────
+  {
+    id: 'answeroverflow',
+    name: 'Answer Overflow',
+    description: 'Search indexed Discord community discussions for coding answers via Answer Overflow.',
+    category: 'productivity',
+    icon: '💬',
+    content: `---
+name: answeroverflow
+description: Search indexed Discord community discussions for coding answers via Answer Overflow
+category: productivity
+icon: 💬
+trigger: discord|answeroverflow|community support
+enabled: true
+source: https://github.com/AnswerOverflow/AnswerOverflow
+---
+
+## What is Answer Overflow?
+Answer Overflow indexes public Discord support channels and makes them searchable via Google and direct API. Perfect for finding answers that only exist in Discord conversations — from servers like Valorant, Cloudflare, C#, Nuxt, and thousands more.
+
+## Quick Search (via Google)
+\`\`\`bash
+# Best approach — Answer Overflow results appear in Google
+web_search "site:answeroverflow.com prisma connection pooling"
+web_search "site:answeroverflow.com nextjs app router error"
+web_search "site:answeroverflow.com discord.js slash commands"
+\`\`\`
+
+## Fetch Thread Content
+\`\`\`bash
+# Markdown format (preferred for agents)
+http_request GET https://www.answeroverflow.com/m/<message-id>
+# Add Accept: text/markdown header, or use /m/ prefix URL
+\`\`\`
+
+URL patterns:
+- Thread: \`https://www.answeroverflow.com/m/<message-id>\`
+- Server: \`https://www.answeroverflow.com/c/<server-slug>\`
+- Channel: \`https://www.answeroverflow.com/c/<server-slug>/<channel-slug>\`
+
+## MCP Server
+Answer Overflow exposes an MCP server at \`https://www.answeroverflow.com/mcp\`:
+
+| Tool | Description |
+|------|-------------|
+| \`search_answeroverflow\` | Search all indexed communities; filter by server/channel ID |
+| \`search_servers\` | Discover indexed Discord servers |
+| \`get_thread_messages\` | Get all messages from a specific thread |
+| \`find_similar_threads\` | Find threads related to a given thread |
+
+To add it as an MCP server in NeoAgent: command \`npx -y @answeroverflow/mcp\`
+
+## Tips
+- Results are real Discord conversations — context may be informal
+- Threads often have back-and-forth before the solution; read the whole thread
+- Check the server/channel name to understand context (official support vs community)
+- Many major open-source projects index their Discord support channels here`
+  },
+
+  // ── FUN ─────────────────────────────────────────────────────────────────────
+  {
+    id: 'random-joke',
+    name: 'Random Joke',
+    description: 'Fetch a random joke (clean, programmer or general).',
+    category: 'fun',
+    icon: '😄',
+    content: `---
+name: random-joke
+description: Fetch a random joke (clean, programmer or general)
+category: fun
+icon: 😄
+enabled: true
+---
+
+GET \`https://v2.jokeapi.dev/joke/Programming,Miscellaneous?blacklistFlags=nsfw,racist,sexist,explicit\`
+
+The response has either a single \`joke\` field or a two-part \`setup\`/\`delivery\`. Present it naturally — pause before the punchline in delivery style. If the user asks for a specific category (dark, pun, etc.), adjust the URL accordingly.`
+  },
+
+  {
+    id: 'random-quote',
+    name: 'Random Quote',
+    description: 'Get a random motivational or philosophical quote.',
+    category: 'fun',
+    icon: '💭',
+    content: `---
+name: random-quote
+description: Get a random motivational or philosophical quote
+category: fun
+icon: 💭
+enabled: true
+---
+
+GET \`https://api.quotable.io/random\`
+
+Show: \`"<content>"\` — *<author>*
+
+If the user specifies a topic or author, use:
+\`https://api.quotable.io/random?tags=<tag>\` (common tags: technology, wisdom, success, life, motivational, literature, science)
+\`https://api.quotable.io/quotes?author=<slug>\` for a specific author.`
+  },
+
+  {
+    id: 'random-fact',
+    name: 'Random Fact',
+    description: 'Get a random interesting fact.',
+    category: 'fun',
+    icon: '🧠',
+    content: `---
+name: random-fact
+description: Get a random interesting fact
+category: fun
+icon: 🧠
+enabled: true
+---
+
+GET \`https://uselessfacts.jsph.pl/api/v2/facts/random?language=en\` for a random fact.
+Alternatively: \`https://api.api-ninjas.com/v1/facts?limit=1\` (no key needed for free tier).
+Present the fact naturally, optionally adding a brief "why this is interesting" comment if it's not immediately obvious.`
+  },
+
+  {
+    id: 'word-definition',
+    name: 'Word Definition',
+    description: 'Look up the definition, pronunciation and examples for any word.',
+    category: 'fun',
+    icon: '📖',
+    content: `---
+name: word-definition
+description: Look up the definition, pronunciation and examples for any word
+category: fun
+icon: 📖
+enabled: true
+---
+
+GET \`https://api.dictionaryapi.dev/api/v2/entries/en/<word>\`
+
+Extract and display:
+- Pronunciation (phonetic)
+- Part of speech
+- Primary definition(s)
+- Example sentence if available
+- Synonyms (first 5)
+
+If the word isn't found, suggest similar spellings. Supports multiple meanings grouped by part of speech.`
+  },
+
+  {
+    id: 'qr-code',
+    name: 'QR Code Generator',
+    description: 'Generate a QR code image URL for any text or URL.',
+    category: 'fun',
+    icon: '⬛',
+    content: `---
+name: qr-code
+description: Generate a QR code image URL for any text or URL
+category: fun
+icon: ⬛
+enabled: true
+---
+
+Use the QR Server API (no auth):
+\`https://api.qrserver.com/v1/create-qr-code/?data=<encoded_text>&size=300x300&margin=10\`
+
+URL-encode the input data. Provide the direct image URL that the user can open in a browser or embed. Also calculate: at the default error correction level (M), the QR can hold the given text reliably up to X characters.`
+  },
+
+  // ── DEV (continued) ──────────────────────────────────────────────────────────
+  {
+    id: 'github',
+    name: 'GitHub',
+    description: 'Interact with GitHub repos, PRs, issues, branches, CI and releases using git and the gh CLI.',
+    category: 'dev',
+    icon: '🐙',
+    content: `---
+name: github
+description: Interact with GitHub repos, PRs, issues, branches, CI and releases using git and the gh CLI
+trigger: When the user asks to clone a repo, create a PR, open/close issues, check CI status, fork, review diffs, manage branches, or do anything GitHub-related
+category: dev
+icon: 🐙
+enabled: true
+---
+
+# GitHub Skill
+
+Use \`git\` for local version control and \`gh\` (GitHub CLI) for GitHub-specific actions.
+Always verify both tools are available: \`which git && which gh\`.
+If \`gh\` is not authenticated, prompt the user to run \`gh auth login\`.
+
+## Repo Status & Info
+\`\`\`
+gh repo view                          # show current repo overview
+gh repo view <owner>/<repo>           # show a specific repo
+git status                            # show working tree state
+git log --oneline -20                 # last 20 commits
+git diff                              # unstaged changes
+git diff --staged                     # staged changes
+\`\`\`
+
+## Clone & Fork
+\`\`\`
+gh repo clone <owner>/<repo>          # clone via gh (sets up remote automatically)
+gh repo fork <owner>/<repo> --clone   # fork and clone in one step
+\`\`\`
+
+## Branches
+\`\`\`
+git checkout -b <branch>              # create and switch to new branch
+git branch -a                         # list all branches (local + remote)
+git push -u origin <branch>           # push new branch to origin
+git branch -d <branch>                # delete local branch
+gh repo sync                          # sync fork with upstream
+\`\`\`
+
+## Commits & Push
+\`\`\`
+git add -A && git commit -m "<msg>"   # stage all and commit
+git push                              # push to tracked remote branch
+git pull --rebase                     # pull with rebase
+\`\`\`
+
+## Pull Requests
+\`\`\`
+gh pr create --title "<title>" --body "<body>"   # open a PR
+gh pr list                                        # list open PRs
+gh pr view <number>                               # view a specific PR
+gh pr checkout <number>                           # check out a PR branch locally
+gh pr merge <number> --squash                     # merge PR (squash)
+gh pr review <number> --approve                   # approve a PR
+gh pr review <number> --request-changes -b "<feedback>"
+gh pr close <number>                              # close without merging
+gh pr status                                      # PRs involving you
+\`\`\`
+
+## Issues
+\`\`\`
+gh issue create --title "<title>" --body "<body>"  # create issue
+gh issue list                                       # list open issues
+gh issue view <number>                              # view an issue
+gh issue close <number>                             # close an issue
+gh issue comment <number> -b "<comment>"            # add a comment
+gh issue assign <number> --assignee @me             # assign to yourself
+\`\`\`
+
+## Releases & Tags
+\`\`\`
+gh release create <tag> --title "<title>" --notes "<notes>"
+gh release list
+gh release view <tag>
+git tag -a v1.0.0 -m "Release v1.0.0" && git push --tags
+\`\`\`
+
+## CI / Actions
+\`\`\`
+gh run list                           # list recent workflow runs
+gh run view <run-id>                  # view a run's details and logs
+gh run watch <run-id>                 # stream live logs
+gh workflow list                      # list workflows
+gh workflow run <workflow-file>       # trigger a workflow manually
+\`\`\`
+
+## Presentation Tips
+- For pr/issue lists, format as a table: number, title, author, date.
+- For git log, show short hash, message, and relative date.
+- For CI runs, summarise: ✅ success / ❌ failure / ⏳ in progress per job.
+- Confirm with the user before any repo-modifying operation (push, merge, delete).`
+  },
+
+  // ── RESEARCH ─────────────────────────────────────────────────────────────────
+  {
+    id: 'deep-research',
+    name: 'Deep Research',
+    description: 'Conduct thorough multi-source research on any topic and synthesise findings into a structured report.',
+    category: 'research',
+    icon: '🔬',
+    content: `---
+name: deep-research
+description: Conduct thorough multi-source research on any topic, synthesising findings into a structured report
+trigger: When the user asks for deep research, a detailed investigation, a comprehensive overview, or wants to understand a complex topic in depth
+category: research
+icon: 🔬
+enabled: true
+---
+
+# Deep Research Skill
+
+Perform iterative, multi-source research by combining web searches, page reads, and cross-referencing. Don't stop at the first result — go wide, then go deep.
+
+## Research Process
+
+### 1. Clarify the Query
+Decompose the topic into 3–5 sub-questions that together fully answer the user's request. State them so the user can correct scope before you begin.
+
+### 2. Broad Discovery (go wide)
+Run multiple searches using varied queries and angles:
+- Direct topic searches: \`<topic> overview\`, \`<topic> explained\`
+- Authoritative sources: \`site:en.wikipedia.org <topic>\`, \`site:arxiv.org <topic>\`
+- Recent developments: \`<topic> 2025\`, \`<topic> latest research\`
+- Opposing views: \`<topic> criticism\`, \`<topic> limitations\`, \`<topic> controversy\`
+
+Use \`https://html.duckduckgo.com/html/?q=<query>\` for web searches (parse \`<a class="result__a">\` links and snippets).
+
+### 3. Deep Reads (go deep)
+For each sub-question, identify the 3–5 most relevant URLs. Fetch and read each one fully. Extract:
+- Key claims and data points
+- Author / publication / date (assess credibility)
+- References to follow up on
+
+### 4. Cross-Reference & Validate
+- Note where sources agree and disagree.
+- Flag claims from only one source as unverified.
+- If numbers or stats are cited, trace to the primary source.
+- Prefer sources from the last 2 years unless historical context is needed.
+
+### 5. Synthesise & Report
+\`\`\`
+## Summary
+2–3 sentence TL;DR.
+
+## Background
+Context needed to understand the topic.
+
+## Key Findings
+- Finding 1 (source: [title](url))
+- Finding 2 ...
+
+## Different Perspectives / Debate
+Where experts or sources disagree, and why.
+
+## Open Questions / Gaps
+What is still unknown or contested.
+
+## Sources
+Numbered list of all URLs consulted, with title and date.
+\`\`\`
+
+## Tips
+- Run at least 5 distinct searches before writing the report.
+- Read at least 4 full pages (not just snippets).
+- Never present a single source as definitive — always triangulate.
+- Cite every factual claim with a link.`
+  },
+
+  {
+    id: 'coding',
+    name: 'Coding',
+    description: 'Write, debug, refactor, explain, and review code in any programming language.',
+    category: 'dev',
+    icon: '💻',
+    content: `---
+name: coding
+description: Write, debug, refactor, explain, and review code in any programming language
+trigger: When the user asks to write code, fix a bug, refactor, explain how code works, review a function, add tests, or help with any programming task
+category: dev
+icon: 💻
+enabled: true
+---
+
+# Coding Skill
+
+Handle the full software development lifecycle: writing new code, understanding existing code, debugging, refactoring, and testing — in any language.
+
+## Write New Code
+- Confirm the language, framework, and environment before starting.
+- Ask for constraints (performance, style guide, existing dependencies).
+- Write clean, idiomatic code with comments for non-obvious logic.
+- Include a usage example or short test block where helpful.
+
+## Debug & Fix
+1. Read the full error message and stack trace carefully.
+2. Identify the exact file and line where the error originates.
+3. Inspect surrounding code for logic errors, type mismatches, null-dereferences, or off-by-one errors.
+4. Run the code in a terminal to reproduce the error if needed.
+5. Apply the minimal fix that resolves the root cause — avoid masking errors with bare try/catch.
+6. Explain what was wrong and why the fix works.
+
+## Refactor
+- Identify code smells: duplication, long functions, deep nesting, magic numbers, unclear naming.
+- Refactor in small, safe steps — preserve behaviour before improving structure.
+- Prefer readability over cleverness.
+- Verify functionality is unchanged after refactoring.
+
+## Explain Code
+- Summarise what the code does at a high level first.
+- Walk through it section by section, explaining intent not just mechanics.
+- Highlight non-obvious patterns (closures, generators, recursion, bitwise tricks, etc.).
+- Point out potential edge cases or bugs noticed during the read.
+
+## Code Review
+Evaluate on these dimensions and provide concrete, actionable feedback:
+- **Correctness** — does it do what it's supposed to? Are edge cases handled?
+- **Readability** — are names clear? Is logic easy to follow?
+- **Performance** — any O(n²) loops, N+1 queries, or memory leaks?
+- **Security** — injection risks, unvalidated input, exposed secrets, insecure defaults?
+- **Tests** — adequate coverage? Are tests meaningful?
+
+## Add Tests
+- Identify the testing framework in use (Jest, pytest, Go test, etc.) or ask.
+- Write unit tests for individual functions, edge cases, and failure paths.
+- Aim for tests that are independent, deterministic, and fast.
+
+## Quick-Run Commands
+\`\`\`
+python3 <file>.py          # Python
+node <file>.js             # Node.js
+npx ts-node <file>.ts      # TypeScript
+go run <file>.go           # Go
+cargo run                  # Rust
+java <ClassName>.java      # Java 11+
+swift <file>.swift         # Swift
+ruby <file>.rb             # Ruby
+\`\`\`
+
+## Pre-Delivery Checklist
+- No hardcoded secrets or credentials
+- Input is validated / sanitised at entry points
+- Error paths are handled (not silently swallowed)
+- Async code uses proper await / error handling
+- No unused imports or dead code left behind`
+  },
+
+  // ── MAKER ────────────────────────────────────────────────────────────────────
+  {
+    id: 'bambu-studio-cli',
+    name: 'BambuStudio CLI',
+    description: 'Slice 3MF/STL files, export G-code and slicing data using BambuStudio on the command line.',
+    category: 'maker',
+    icon: '🖨️',
+    content: `---
+name: bambu-studio-cli
+description: Slice 3MF/STL files and export results using the BambuStudio command-line interface
+category: maker
+icon: 🖨️
+enabled: true
+---
+
+# BambuStudio CLI
+
+Invoke BambuStudio headlessly for slicing and exporting. The binary is typically \`bambu-studio\` on Linux/macOS or \`bambu-studio.exe\` on Windows.
+
+## Core flags
+
+| Flag | Description |
+|------|-------------|
+| \`--slice <plate>\` | Slice plates: \`0\` = all, \`N\` = plate N |
+| \`--export-3mf <out.3mf>\` | Export sliced result as 3MF |
+| \`--outputdir <dir>\` | Directory for all exported files |
+| \`--load-settings "machine.json;process.json"\` | Override printer + process settings |
+| \`--load-filaments "f1.json;f2.json"\` | Override filament settings (use \`;\` separators, skip slots with empty entry) |
+| \`--curr-bed-type "Cool Plate"\` | Set bed type via command line |
+| \`--arrange <0\|1>\` | Auto-arrange: 0=off, 1=on |
+| \`--orient\` | Auto-orient models before slicing |
+| \`--scale <factor>\` | Scale model by float factor (e.g. \`1.5\`) |
+| \`--export-settings <out.json>\` | Dump merged settings to JSON |
+| \`--export-slicedata <dir>\` | Export slicing data to folder |
+| \`--load-slicedata <dir>\` | Load cached slicing data |
+| \`--info\` | Print model info without slicing |
+| \`--debug <0-5>\` | Log level: 0=fatal … 5=trace |
+| \`--pipe <name>\` | Send progress to named pipe |
+| \`--uptodate\` | Upgrade 3MF config values to latest profiles |
+| \`--help\` | Show CLI help |
+
+Setting priority (highest → lowest):
+1. \`--key=value\` flags on the command line
+2. Files loaded via \`--load-settings\` / \`--load-filaments\`
+3. Settings embedded in the 3MF file
+
+## Common usage patterns
+
+### Slice a 3MF using its own settings
+\`\`\`bash
+bambu-studio --slice 0 --debug 2 --export-3mf output.3mf model.3mf
+\`\`\`
+Slices all plates in model.3mf and exports to output.3mf.
+
+### Slice a 3MF with custom machine/process/filament overrides
+\`\`\`bash
+bambu-studio \\
+  --load-settings "machine.json;process.json" \\
+  --load-filaments "filament1.json;;filament3.json" \\
+  --curr-bed-type "Cool Plate" \\
+  --slice 2 --debug 2 \\
+  --export-3mf output.3mf \\
+  model.3mf
+\`\`\`
+Slices plate 2 only, overriding printer/process/filament settings from JSON files.
+Empty \`;\;\` entries keep the filament slot from the 3MF unchanged.
+
+### Slice raw STL files
+\`\`\`bash
+bambu-studio \\
+  --orient --arrange 1 \\
+  --load-settings "machine.json;process.json" \\
+  --load-filaments "filament.json" \\
+  --slice 0 --debug 2 \\
+  --export-3mf output.3mf \\
+  model.stl
+\`\`\`
+Auto-orients and arranges the STL, applies settings from JSON files, slices all plates.
+
+## When the user asks to slice a file:
+1. Confirm the path to \`bambu-studio\` binary (run \`which bambu-studio\` or ask the user)
+2. Check if custom settings JSON files are needed or if the 3MF is self-contained
+3. Build the command from the flags above
+4. Run it and check for errors in the output (debug level 2 is a good default)
+5. Report the output file location and any warnings`
+  }
+];
+
+// ── Routes ─────────────────────────────────────────────────────────────────────
+
+/** GET /api/store — return catalog with installed status */
+router.get('/', (req, res) => {
+  const fs = require('fs');
+  const installed = new Set();
+
+  if (fs.existsSync(SKILLS_DIR)) {
+    for (const f of fs.readdirSync(SKILLS_DIR)) {
+      installed.add(f.replace(/\.md$/i, ''));
+    }
+  }
+
+  const items = CATALOG.map(s => ({
+    id: s.id,
+    name: s.name,
+    description: s.description,
+    category: s.category,
+    icon: s.icon,
+    installed: installed.has(s.id)
+  }));
+
+  res.json(items);
+});
+
+/** POST /api/store/:id/install — write the skill file */
+router.post('/:id/install', (req, res) => {
+  const fs = require('fs');
+  const skill = CATALOG.find(s => s.id === req.params.id);
+  if (!skill) return res.status(404).json({ error: 'Skill not found in catalog' });
+
+  if (!fs.existsSync(SKILLS_DIR)) fs.mkdirSync(SKILLS_DIR, { recursive: true });
+
+  const filePath = path.join(SKILLS_DIR, `${skill.id}.md`);
+  fs.writeFileSync(filePath, skill.content, 'utf-8');
+
+  // Also reload the skill runner if available
+  const skillRunner = req.app.locals?.skillRunner;
+  if (skillRunner) skillRunner.loadSkillFile(filePath);
+
+  res.json({ success: true, id: skill.id, name: skill.name, filePath });
+});
+
+/** DELETE /api/store/:id/uninstall — remove the skill file */
+router.delete('/:id/uninstall', (req, res) => {
+  const fs = require('fs');
+  const skill = CATALOG.find(s => s.id === req.params.id);
+  if (!skill) return res.status(404).json({ error: 'Skill not found in catalog' });
+
+  const filePath = path.join(SKILLS_DIR, `${skill.id}.md`);
+  if (fs.existsSync(filePath)) fs.unlinkSync(filePath);
+
+  res.json({ success: true, id: skill.id });
+});
+
+module.exports = router;