Today's drop: 2 new tools, 3 listing updates
The directory now lists 120 hand-checked tools. New today, both for IT and security: 1Password for Agents, which lets an AI agent sign in for you while the password stays in your vault; and CVEngine, a free, self-hosted attack-surface scanner that matches what you expose to the internet against CVEs attackers are actually exploiting. Updated: Replit (Monitoring Agent), Canva (Canva Craft image model) and Lovable (Supabase projects created for you). AI for IT & Cybersecurity grows to 37 tools. Pricing only where the official page states it.
Previous drop: Oct 26, 2026 →
Get the weekly dropThe best new AI tools, one email a week. Free.
New in the directory 2
Browse all 120Tool
Category
Price
Status
Link
1Password for Agents1Password's sign-in layer for personal AI agents, launched Oct 22, 2026. When an agent reaches a login page it asks 1Password for the credential; you see which login it wants and approve or deny on your own device, and 1Password fills the password or one-time code straight into the site, so the agent never sees or stores it. You can disconnect an agent at any time. Integrations for Meta's Muse and Grok Bot are rolling out first; Anthropic's Claude Duet, OpenAI's ChatGPT Work and dots and Town's Townie are listed as coming soon, with Browserbase, Browser Use, Perplexity and others planned. 1Password.com accounts only (not 1Password.eu or .ca), and safeguards vary by integration.
Price No extra cost with 1Password Individual and Families plans
CVEngineFree, self-hosted external attack-surface scanner with optional AI triage. Add a domain, IP range or URL and it discovers subdomains (bbot), sweeps ports (naabu), fingerprints services (nmap), runs nuclei checks and matches what you run against exploited CVEs from CISA KEV, EPSS, Exploit-DB, Metasploit and other feeds, then confirms what it can with read-only checks that record the exact command they ran. Bring your own AI (Anthropic, OpenAI, OpenRouter, or local Ollama or vLLM) or run it without one; AI-drafted nuclei templates never run until a person approves them. Includes an MCP server and REST API with read-only tokens for agents. One Docker Compose stack on Linux (4GB RAM minimum, 8GB recommended). A fork of reNgine 2.x from a single maintainer, published Oct 24, 2026 and at about 400 GitHub stars two days later. Only scan systems you own or are authorized to test.
Price Open source (GPL-3.0)
Why these two
- 1Password for Agents answers the question every agent user hits sooner or later: what do I do at the login page? Pasting a password into a prompt hands it to the agent for good. With this, the agent asks 1Password for a specific login, you see which one and approve or deny it on your own device, and 1Password fills the password or one-time code into the site itself. The agent only learns that it got in. It launched Oct 22 at no extra cost for 1Password Individual and Families accounts on 1Password.com. Be realistic about timing: the Muse and Grok Bot integrations are rolling out first, while Claude Duet, ChatGPT Work and dots and Townie are listed as coming soon, and 1Password says safeguards vary by integration. It pairs with Monday's Strands Box: Box keeps API keys away from coding agents, 1Password does the same for the website logins personal agents use.
- CVEngine is the security repo of the weekend: published Oct 24 and at about 400 GitHub stars two days later, with four releases on Oct 26 alone. Give it a domain or IP range you own and it runs the whole external-assessment loop: subdomain discovery with bbot, port sweeps with naabu, service fingerprinting with nmap, nuclei checks, then matching against exploited-CVE sources such as CISA KEV, EPSS, Exploit-DB and Metasploit. It confirms what it can with read-only checks and records the exact command each one ran. AI is optional (your own Anthropic, OpenAI or OpenRouter key, or a local model through Ollama or vLLM), and any nuclei template the AI drafts waits for a person to approve it. Caveats: it's a brand-new fork of reNgine 2.x from one maintainer, it publishes no accuracy or false-positive numbers, and you need a Linux box with Docker and at least 4GB of RAM. Only scan what you own or are authorized to test.
Listing updates
Checked on Oct 27, 2026 against official pages and announcements.
- Replit: Monitoring Agent (announced Oct 22, in the Oct 23 changelog) is an on-call agent for published apps. With monitoring alerts and Auto-fix on, it investigates production issues from your app's logs and metrics, attempts and validates a fix, and emails you a summary. The experimental Auto-republish setting lets it publish the fix too; leave it off to review each change yourself. Replit hasn't published accuracy numbers.
- Canva: Canva Craft (Oct 22) is Canva's first in-house image model, built by Canva Research and available in Dream Lab to all Canva users. It plans the composition before rendering, makes images up to 2K in more than 30 aspect ratios with text in 55 languages, and its output converts to editable layers with Magic Layers. Developers can call it through Canva's preview Connect API as
craft_v1_2k(about 50-60 seconds per job, using the user's AI credits); Canva warns that preview APIs can change without notice. - Lovable: since Oct 20 you can ask Lovable to create a Supabase project for an app with no backend. You pick the organization and region (the region can't be changed later), and Supabase bills that project to your own organization. Companies running several Lovable workspaces can also group them into an organization with SSO and SCIM set up once.
Category changes
- AI for IT & Cybersecurity grows from 35 to 37 tools. The Breaking into IT guide adds CVEngine (self-host it on a lab VM and learn the commands behind each finding) and 1Password for Agents (approve each login instead of handing your agent a password).
Checked, not listed
- zCAD (prompt-to-parametric CAD with STEP export, a Product Hunt #1): a real product, but mechanical CAD is outside this directory's audience, its AI side is in beta, and free-plan models are public by default.
- AgentValley (a free, MIT-licensed Mac notch monitor for Claude Code sessions): neat, but about 8 GitHub stars so far. We'll check back.
- browserpilot (local browser AI that reads CRMs via OCR) and CouncilAI (a Windows installer for Karpathy's LLM Council): each is a single Reddit post with no traction we could verify yet.
- Cisco MCP Scanner (about 1,100 GitHub stars): solid, but it overlaps Snyk Agent Scan and Cisco Skill Scanner, which we already list.