Sploitly

Security alerts for AI-built apps

Know when something real hits your stack. Let your agent fix it.

Sploitly watches the packages your app runs on. When attackers start using a flaw in one of them, your coding agent tells you at the start of your next session, with one line it can act on. Less noise, not more.

Example alertmy_alerts
project
shop-app
package
lodash (npm)
installed
4.17.20
fixed in
4.17.21
why now
rated high
Paste to agent
Fix Sploitly alert 123: upgrade lodash from 4.17.20 to 4.17.21 in project shop-app

What it does

It watches your stack, not the headlines

Every scary security headline is not your problem. Sploitly checks the news against the exact versions in your lockfile and only speaks up when it matters to you.

Watches public sources

The CISA Known Exploited Vulnerabilities list, public security advisories, and whether a free attack tool exists (Metasploit, Nuclei, Exploit-DB).

Checks your saved packages daily

Your saved list is re-checked every day. A new problem becomes an alert: when attackers are using it, when a free attack tool exists, or when it is rated high or critical.

Hands your agent the fix

Each alert names the project, package, installed version and fixed version, plus one line to paste to your agent. Save the updated list and the alert closes by itself.

Or posts it to a webhook

Give us one HTTPS address and each new alert arrives as signed JSON. Route it anywhere from there with your own tools.

How it works

Three steps, then it runs itself

  1. Connect your agent

    You get a setup code like SP-7K3M-9QXA-2HDF. It works once and expires after 24 hours, so it is safe to paste anywhere. One-line setup connects Sploitly to the coding agents on your machine. Coming at launch

    Agents can also read the plain-text guide at /start.

  2. Approve your package list

    Your agent reads your lockfile and builds a list: package names, exact versions, ecosystem. It shows you the list, and it is told to save it only after you approve.

  3. Hear about what matters

    At the start of each session your agent calls my_alerts and tells you about anything open. Before adding or upgrading a package it calls check_packages first.

Which agents

Any agent that speaks MCP

Sploitly is an MCP server, so it works with any coding agent that supports MCP. We have tested it with these:

  • Claude Code
  • Cursor
  • Codex

No MCP support? Your agent or your own scripts can use the plain HTTPS API instead.

Your data

What we store, and what we never store

We store

  • Package names, exact versions and ecosystem, from your lockfile
  • The project label you choose
  • Your webhook address, if you set one, and the status code of each delivery

We never store

  • Your code
  • Secrets or API keys from your project
  • File contents or file paths

Guard rails

  • Saving a watch list needs your approval. Enforced at launch
  • Your agent never needs to see your key: connections sign in through the browser. Coming at launch
  • Deleting a project's list is not an agent tool, on purpose, so a poisoned web page cannot talk your agent into it.

Prefer not to save a list? Your agent can run check_packages whenever you like instead. Details on the privacy page.

FAQ

Questions

Will this flood me with alerts?

No. You get an alert only when attackers are using the flaw, when a free attack tool exists, or when it is rated high or critical. Lower ones still show up when your agent runs check_packages, they just do not ping you.

Do you see my code?

No. We get package names, exact versions and the ecosystem. Never code, secrets, file contents or paths.

Which package ecosystems are covered?

npm, PyPI, Go, crates.io, RubyGems, Maven, NuGet and Packagist.

How do I know an alert is really fixed?

Upgrade, then save your package list again. If the package is no longer affected, the alert closes by itself.

Can alerts go to Slack, email or my phone?

Alerts reach you through your agent, or through one signed webhook you control. From the webhook you can route them anywhere with your own tools, such as Zapier or n8n. Email alerts are not available yet.

Can I delete my data?

Yes, at any time, through the API. Deleting a project's list also closes its alerts.

What does it cost?

Pricing is announced at launch. See the plans.

Build fast. Hear about the attacks that matter.

Sploitly is in early access. Tell us what you are building and which agent you use.