Cve Intel

Every known software vulnerability gets a public ID, such as CVE-2024-3400. For each one, security, IT and development teams face the same question: does this affect us, and do we fix it now or can it wait? Answering it usually means checking NVD, CISA, FIRST, GitHub and vendor advisories by hand, one browser tab at a time.

Why it matters now

CVE publication keeps rising (48,185 in 2025, 20.6% more than 2024, and over 86,000 by 27 October 2026), and in April 2026 NIST said the National Vulnerability Database cannot keep up, so it will enrich CVEs in CISA's Known Exploited Vulnerabilities (KEV) catalog, federal-use software and critical software first, and move its unenriched backlog from before 1 March 2026 to "Not Scheduled". Meanwhile, Mandiant's M-Trends 2026 estimates the mean time to exploit at −7 days, so exploitation routinely begins before a patch is released, and exploits were the most common initial infection vector for the sixth year running (32% of intrusions).

What it does

Give CVE Intel a CVE ID, a project folder, a lockfile or SBOM, a container image, the software on this computer, your own domains, or a suspicious IP address, file hash or domain. It queries up to 40 public sources, including CISA KEV, FIRST's EPSS and ENISA's EU Vulnerability Database alongside NVD, and returns what the vulnerability is, a risk score from 0 to 100, whether attackers are already using it, the evidence behind each signal and what to do next. It also shows what changed this week across those sources, and watches the vendors, products and packages you run.

• 40 data sources for vulnerability, exploit, network and threat intelligence, listed under Data sources. Many need no API key.
• 41 MCP tools for AI agents and coding assistants. The same program is also a command line and a REST API (see Developer notes).
• SBOMs in and out. It reads CycloneDX and SPDX SBOMs along with lockfiles, and Save SBOM (or scan deps . --format cyclonedx) writes any scan as a CycloneDX 1.5 SBOM with the findings, their risk signals and your recorded decisions.
• Local. There is no CVE Intel account or server. Lookups go straight from your computer to the public sources, and the AI assistant runs on a local model unless you choose a cloud provider.

It is for security analysts, IT and operations teams, developers checking their dependencies, and anyone learning vulnerability management.

What it helps you do

• Decide whether to patch now. Enter a CVE ID and get a risk score from 0 to 100, a plain recommendation and how urgent it is, with the evidence behind it.
• Put a list of vulnerabilities in order. Paste up to 10 CVE IDs from a scanner report and get the order to patch them in.
• Check your project's dependencies. Drop a project folder on the window, or paste a lockfile or SBOM. You see which packages have known vulnerabilities, ranked by how likely they are to hurt you (Log4Shell first, not just the highest CVSS), and which version fixes them. Scan again later to see what is new, fixed or riskier since last time, or save the scan as a CycloneDX SBOM with the findings in it.
• Check what you run and ship. Scan the apps and operating system on this computer, or a container image you build (no Docker needed). Or see what public sources already know about your own domains and servers, without CVE Intel ever connecting to them.
• See what affects the products you run. List the known vulnerabilities for a product and version, such as Apache Tomcat 9.0.80, with actively exploited ones first.
• See what changed this week. Open Landscape for new CISA KEV entries, CVEs ENISA marks exploited that KEV does not list yet, rising EPSS scores, new Metasploit modules and PoCs, new critical CVEs, Project Zero's in-the-wild zero-days and the latest CISA advisories. CVEs that several sources agree on are listed first.
• Watch the products you run. Add a vendor, product, package or CVE to the local watchlist. Each refresh records what changed: new CVEs, KEV listings, EPSS jumps, new exploits, CVSS revisions and newly available fixes.
• Remember what you decided, and hear when the reasons change. Record a decision on a CVE (patch now, defer until a date, accept the risk, not affected, mitigated) with the assumptions it rests on. Every refresh re-checks them: a deferral turns stale the day the CVE enters CISA KEV or an exploit appears, and the ledger tells you how your past calls held up.
• Keep an eye on what attackers are exploiting. Filter CISA's catalog of actively exploited vulnerabilities by vendor, product, ransomware use or date.
• Investigate something suspicious. Check an IP address from your logs, a file hash your antivirus flagged, or an unfamiliar domain or URL against threat-intelligence sources.
• Ask questions in plain English. "Should we patch CVE-2024-3400 this weekend?" The AI assistant looks up the facts first, then explains them, and shows which sources it used.
• Give AI coding assistants the same knowledge. Claude Code, Cursor, VS Code and other tools can use CVE Intel while they work, through MCP.

Terms you'll see

The risk score combines these signals: severity (CVSS), the chance of exploitation (EPSS), confirmed exploitation (CISA KEV) and public exploit code. Anything on CISA KEV is always CRITICAL. See Risk score for the formula.

CVE Intel helps you decide what to patch. It does not probe, port-scan or attack anything, and it does not apply fixes. Scans read what is already on your machine (dependency files, the software installed on this computer, a container image you choose) and send only package and product names and versions to OSV.dev and NVD. The exposure check is passive: it asks public sources (Shodan, certificate transparency logs) what they already know about domains and addresses you own, and sends nothing to them. Results are only as current as the public sources.

Getting started

You don't need an account or any API keys to start. Most people are looking at their first result within a couple of minutes:

1. Download the app for your operating system and open it. The Landscape view loads first: what changed this week across the public sources.
2. Open CVE Triage, type a CVE ID such as CVE-2024-3400 and press Triage, or click one of the examples under the search box. Press Watch CVE (or Watch next to the vendor) to be told when it changes.
3. Drag a project folder onto the window to check its dependencies, or open Scan and choose This computer to check the software installed on this machine. Findings come back ranked by real-world risk, with the version that fixes each one.
4. Press ⌘K (Ctrl+K on Windows and Linux) to open Quick Find. Paste anything there (a CVE, IP address, file hash, domain, URL or package:ecosystem:version) and CVE Intel picks the right lookup.
5. Optional: in Settings → Data source API keys, add free API keys for faster lookups and more exploit data. See API keys.
6. Optional: open Local Model and download a model to use Ask AI with answers that never leave your computer.
7. Optional: open MCP Server to connect your AI coding assistant.

The buttons at the bottom of the sidebar switch between the light and dark themes and make the text larger or smaller. The version you are running is shown below them.

Everything up to Developer notes is about the desktop app. To use CVE Intel from a terminal, a script, a CI pipeline or another program, go straight to that section.

Features

Threat landscape

What changed recently, from several independent sources at once: new CISA KEV entries (with ransomware flags and vendor counts), CVEs ENISA EUVD marks exploited that CISA KEV does not list, the largest EPSS rises, new or updated Metasploit modules and PoC-in-GitHub repositories, newly published critical NVD CVEs, Project Zero's in-the-wild zero-days and the latest CISA advisories. Each panel names its source and when it was fetched, and the ⓘ on each tile explains what it counts. Your watchlist comes first, then the CVEs that appear in two or more panels. Each watched item is a chip that opens it in the Watchlist. Copy the briefing as Markdown for a ticket or weekly note. Ask "What's new this week?" in Ask AI for the same briefing.

Watchlist

Watch the vendors, products, packages and CVEs you run. Each refresh compares them with the last check and records what changed: new CVEs, KEV listings, ransomware use, EPSS jumps, new exploits or Metasploit modules, CVSS changes, newly available fixes and ENISA exploitation reports. The first check of an item records a silent baseline. The list lives only on this machine (~/.cve-intel/watch.db, owner-only). While the app is open it can refresh on a schedule (every 6 hours by default; change or turn it off in Settings). AI agents can read the watchlist over MCP but cannot change it.

Decisions

Every tool says what is risky. None remembers what you decided, why, and whether the reasons still hold. Decisions does. Press Record decision on a triage result, a scan finding or a comparison, choose a verdict (patch now, defer until a date, accept the risk, not affected, mitigated) and tick the assumptions it rests on. The machine-checked ones are pre-ticked from today's evidence: not in CISA KEV, EPSS below a threshold, no public exploit, no Metasploit module, local access only, needs privileges, needs user interaction, no fix yet, not marked exploited by ENISA, no ransomware use, CVSS below a threshold, the sources agree, NVD has not analysed it yet. The manual ones (not internet-facing, behind a WAF, a compensating control, not reachable, the vendor says not affected, or your own words) carry a review date.

Recording a decision adds the CVE to the watchlist, and every refresh re-checks the assumptions. The day the CVE enters KEV, EPSS crosses the threshold, an exploit or a fix appears, or NVD publishes its analysis, the decision turns stale with the exact reason, an event lands in the watchlist feed, and the Landscape counts it under Decisions to revisit. A deferral past its date is expired; a manual assumption past its review date is review due. Re-decide replaces a decision and keeps the old one as history; Close records whether it was patched, accepted for good, won't be fixed, or other.

The ledger also tells you how your judgment held up: of the CVEs you held back in the last 30, 90 or 365 days, how many later entered KEV, gained an exploit or rose in EPSS, and how many days after the decision; of the patch-now calls, how many were confirmed by KEV, an exploit or a high EPSS at the time or since. Export the ledger as OpenVEX or CycloneDX VEX for scanners and SBOM consumers (the CycloneDX document lists each decided-on package as a component, so its references resolve in Dependency-Track and the like). It lives only on this machine (~/.cve-intel/decisions.db, owner-only); AI agents can read it over MCP (list_decisions, get_decision_calibration) but never change it, and Ask AI mentions your decision when you ask about a CVE you decided on. From the terminal: cve-intel decide record CVE-2025-12345 --verdict defer --until 2026-12-31 --auto, then decide list, decide check, decide calibration and decide export.

CVE Triage

One lookup across NVD, CVE.org, EPSS, CISA KEV, ENISA EUVD and public exploit sources. The result includes the risk score, a plain recommendation and urgency, and the evidence behind it: CVSS (and where it came from), 30-day EPSS probability and its recent trend, KEV dates, ENISA's exploited flag, CIRCL sightings (deep triage), and the exploit evidence found in Metasploit, Exploit-DB, Nuclei, GitHub, PoC-in-GitHub and, with a VulnCheck token, VulnCheck KEV and XDB. When NVD has not scored a CVE yet, the CNA's or CISA's score from the CVE.org record is used. Deep triage also returns a CISA SSVC decision, using CISA's own published SSVC assessment (Vulnrichment) where one exists. ENISA EUVD or VulnCheck KEV can raise SSVC Exploitation to active; they do not change the risk score.

Who says what. Every triage result also lays the sources side by side: NVD's own score and analysis status, the CNA's score, CISA's ADP (Vulnrichment) score and SSVC assessment, EPSS, CISA KEV, ENISA EUVD, VulnCheck KEV and, in deep triage, Microsoft, Red Hat and Ubuntu (press Check vendors on a standard triage to add them). It names where they disagree: scores 2 or more points apart, a CVE that NVD has not analysed (common since NVD limited its analysis in 2026), a critical score that attackers have ignored for six months, exploitation reported by a source other than CISA KEV, a distribution that says its build is not affected or will not be fixed, and rejected or disputed CVEs. The card is marked Sources agree, Sources disagree or Thin evidence, and the same appears as the Agreement row in Compare CVEs, in the consensus field of triage_cve, and in the CLI's triage output. None of it changes the risk score.

Exploit speed. EPSS is a daily statistical model and the exploit indexes record what already exists; both lag. With language models reading patches, the public fix commit of an open-source CVE is itself a leading signal. A standard or deep triage therefore shows an Exploit speed estimate: Hours, Days, Weeks, Unknown, or Exploit exists when a PoC or a KEV listing already makes the question moot. It starts at 30 points and adds or subtracts for the fix (a small, localised public fix with a test that encodes the trigger counts most; no public fix at all counts against), the weakness class (injection, traversal, authentication bypass and deserialisation are fast; memory corruption and races are slow), the CVSS vector (network, no privileges, no user interaction) and the time already elapsed without a PoC. Every line of the rationale is shown, the label says estimate, the confidence never exceeds medium, and without a public fix the label is capped at Days. Fix commits come from the references NVD, CVE.org, OSV and GitHub Advisories tag as the patch and are read from GitHub (two requests per CVE at most; a free token raises the limit from 60 to 5,000 an hour). The same estimate is a row in Compare CVEs, a chip on the riskiest unexploited findings of a scan, the exploit_speed field of triage_cve and scan_dependencies, and the fix date on get_cve_timeline. It never changes the risk score.

The interface follows your system's light or dark theme until you pick one with the sun and moon button at the bottom of the sidebar. Beside it, A− and A+ make the text and controls smaller or larger in every view (⌘ −, ⌘ + and ⌘ 0 on macOS; Ctrl on Windows and Linux), and the size is remembered.

Ask in plain language

Ask questions such as "Which is worse: CVE-2024-3400 or CVE-2023-44487?" CVE Intel runs the lookups first, then the model explains the results and lists the tools it used. Small local models can still get details wrong, so check an answer you will act on against CVE Triage.

Compare and prioritise

Paste 2 to 10 CVE IDs to compare them and get a patch order based on KEV status, score, EPSS and exploit evidence, not on CVSS alone. Not sure where to start? The examples under the box (KEV vs. CVSS, VPN & firewall zero-days, Microsoft patch day and Open-source libraries) each show one way the order can surprise you.

CISA KEV CVEs come first, then the risk score. Equal scores are common (exploited CVEs with a high EPSS all reach 100), so ties are broken by known ransomware use, EPSS, exploit maturity, CVSS and the earlier KEV due date. Each CVE says why it sits above the next one.

• Side by side: one column per CVE, with rows for the score's parts, KEV dates, ransomware use, who reports exploitation, EPSS and its 30-day trend, exploit sources, CVSS attack vector, complexity, privileges, user interaction and impact, product, CWE and age. Riskier values that differ are in red. Only rows that differ hides the rest.
• What they have in common: CVEs in the same product (one update may cover them) or from the same vendor, ransomware use, exploitation reported outside CISA KEV (which the score doesn't count), network access without sign-in or user action, public exploits, rising EPSS, and shared weaknesses.

Scan dependencies

Drop a project folder or dependency files anywhere on the window, or use Choose folder… / Choose files…. A folder is searched, including subfolders, for the lockfiles, manifests and SBOMs listed below. The search skips node_modules, virtualenvs, vendor, build output and hidden folders, and skips a package.json or go.sum that sits next to the lockfile that pins it. Each file found is listed with its format and package count. Untick any you don't want, then press Scan. Every file is parsed on its own, a package listed in several files is checked once, and each finding names the files it came from. Files are read on your machine; only package names and versions are sent to OSV.dev.

You can also paste a lockfile (package-lock.json, yarn.lock, Pipfile.lock, poetry.lock, uv.lock, Cargo.lock, go.sum, go.mod, composer.lock, Gemfile.lock), an SBOM (CycloneDX JSON or XML, SPDX JSON or tag-value), a requirements.txt, package.json or pom.xml, purls, or name:ecosystem:version lines (for example from apk info -v or dpkg -l; Alpine needs its release, as in openssl:Alpine:v3.19:3.1.4-r0). The format is detected automatically and up to 10,000 packages are checked against OSV.dev.

Findings are ranked by risk, not just severity. Each one gets the same signals as CVE Triage: whether its CVE is in CISA KEV (and used by ransomware), its EPSS score, exploit evidence from Exploit-DB and Metasploit, and the risk score. Log4Shell in an old log4j-core comes out on top; a CVSS 9.8 nobody exploits does not. When GitHub, PyPI and a distribution each publish the same issue, it is shown once. Filter by risk, KEV or whether a fix exists, or switch to By package to see each vulnerable package with its advisories and the version that fixes them. In a CI pipeline the same scan can fail the build or feed GitHub code scanning; see CI and automation.

Every scan is saved in History, on this computer only. Scan the same folder again (or press Scan again) and the results say what changed since the scan before: findings that are new, resolved (and whether an upgrade or a removal fixed them), or riskier because they were added to CISA KEV, gained an exploit or jumped in EPSS. A pasted list is saved under a hash of its packages, never its text. Copy as Markdown puts the findings table on the clipboard for a ticket, and Forget deletes every scan of a target.

Save SBOM writes the scan you just ran as a CycloneDX 1.5 document: every package or app it read, clean ones included, each finding as a vulnerability affecting its component with the risk signals (KEV, EPSS, exploit evidence, risk score) and the fix version, and your recorded decision on it as the analysis. It is a scan-derived inventory (what the lockfiles, image or this computer declared, with no supplier, hashes, licences or dependency graph), and says so in its metadata; a build-time generator such as Syft or cdxgen remains the authoritative source for those. Load it into Dependency-Track, attach it to a release, or hand it to an auditor. The same document comes from cve-intel scan deps . --format cyclonedx on the command line.

Chains. A ranked list hides chains: a LOW local privilege escalation and a HIGH network RCE on the same host are a full compromise together. The Chains tab of every scan links the findings of a target into attack chains, from network entry through code execution, privilege escalation, credential access or persistence to an outcome. Each step names its basis: the CVSS vector (network, no privileges, scope change), the weakness, CTID's curated KEV → ATT&CK mapping, or an inference from the weakness through CAPEC. Chains get a transparent score (stages covered, the riskiest step, an exploited entry, an unfixed privilege escalation) that never touches a finding's risk score, and the callout at the top names the break point: the finding whose fix cuts the most chains, which may well be that LOW. On this computer and in a container image a chain may start at code execution, since code already runs there. Every result states its assumptions: all findings of a target are treated as present together, and nothing is known about reachability or mitigations. The chains are in the scan's Markdown, in scan … --chains and scans show … --chains on the command line, and in the read-only get_scan_chains tool for local agents.

Watch all affected adds the affected packages and apps of a scan to the Watchlist, riskiest first, so you hear when one of them is added to CISA KEV or gains an exploit. Tick Monitor next to a target in History to have CVE Intel scan it again every day while the app is open (the same folders, this computer, the same image or the same exposure targets); what becomes new or riskier appears in the Watchlist feed and on the Landscape.

Scan this computer

Scan → This computer checks the operating system and the apps installed on this machine. On Windows they come from the registry (the programs in Installed apps, including per-user installs) and the Windows build; on macOS from Applications, ~/Applications, Homebrew (Apple silicon and Intel) and the macOS version. List installed software shows what was found and what will be checked, without contacting anything. Check sends NVD only the vendor, product and version of the apps CVE Intel recognises (browsers, Java, Acrobat, archivers, remote-access tools, VPN clients, developer tools and about 100 more) and of the OS, then ranks the findings like any other scan.

Apps CVE Intel does not recognise are listed as not checked, with the reason. Tick Also look up apps CVE Intel does not recognise to ask NVD's product dictionary for them: this sends their names to NVD, and a match is kept only when NVD's vendor fits the app's publisher. Without an NVD API key NVD allows a few requests a minute, so a first scan checks the riskiest apps (browsers, Java, Acrobat, the OS) first and leaves the rest for the next scan; answers are cached, so later scans are quicker. Programs that are parts of others (system components, drivers) are skipped, and Microsoft 365 and the Visual C++ runtimes are reported as not checked because NVD lists them without versions.

Scan a container image

Scan → Container image checks a saved image without Docker: drop a docker save or OCI archive (.tar, .tar.gz) on the window, or choose an OCI layout folder. The image is read on your computer in one pass, layer by layer, without unpacking anything to disk. Its operating system packages (Debian, Ubuntu, Alpine, Wolfi, Chainguard, Red Hat, Rocky, AlmaLinux, Azure Linux and openSUSE) and any lockfiles, manifests or SBOMs inside it are checked with OSV.dev and ranked like any other scan. A multi-platform image is scanned for this computer's architecture unless you name another platform. Images whose distribution OSV does not cover (Fedora, Amazon Linux, SUSE Linux Enterprise, Debian testing) and older rpm databases are listed as not checked rather than reported clean.

docker save myapp:latest -o myapp.tar
cve-intel scan image myapp.tar --fail-on kev

Check your public exposure

Scan → Public exposure shows what public sources already know about domains and addresses you own. Enter domains or public IP addresses, confirm they are yours (or that you are authorised to assess them), and press Find addresses: CVE Intel collects host names from certificate transparency logs (crt.sh, keeping only names under your domain), resolves them, and lists the public addresses, dropping private and reserved ones. Untick any that are not yours, then press Check. For each address it asks Shodan (with a Shodan key) or the free Shodan InternetDB what services, software versions and CVEs have been seen there, then checks the versions with NVD and ranks the findings like any other scan.

The check is passive: CVE Intel never connects to the targets, and nothing is port-scanned. Addresses that belong to a CDN or hosting platform are flagged and left out, because the software seen there is the provider's. CVEs Shodan infers from a banner are marked as reported, not verified, and distribution builds (Ubuntu, Debian, RHEL) may have the fix backported.

cve-intel scan exposure example.com --authorized --list-only   # names and addresses only
cve-intel scan exposure example.com 203.0.113.10 --authorized --max-ips 20

Databases and data sources

See every upstream source: what it provides, whether it needs a key, whether your key is set, how long results are cached, and which tools use it. Press Check connectivity to probe all of them live. The view also shows the local cache, downloaded models and the model runtime.

MCP server for AI agents

Give AI agents CVE Intel's 41 tools while they work. Start the server from the app, choose your agent, and copy setup steps with the correct address and executable path already filled in. Guides are included for Claude Code, Cursor, VS Code, Claude Desktop, Windsurf and any other MCP client.

• Two ways to connect. The agent can connect over HTTP to the server the app runs, or launch its own copy over stdio. The stdio option works even when the app is closed.
• Simple control. Start and stop the server, pick a port (8000 by default), and optionally start it whenever the app opens. The sidebar shows whether it is running.
• Local only. The server listens on 127.0.0.1 and refuses requests from web pages, so websites and other computers cannot use it.
• Uses your saved keys. Agents get the same data sources and API keys you configured in Settings.

To configure an agent by hand instead, see MCP server setup.

Also included

• Quick Find (⌘K or Ctrl+K): paste a CVE, IP, hash, domain, URL or package:ecosystem:version and CVE Intel picks the right lookup.
• Fixed versions: the version to upgrade a package to so that every known vulnerability is fixed, re-checked against OSV. Quick Find lodash:npm:4.17.15 to see it, or ask in Ask AI.
• Product and KEV search: every CVE affecting a product and version, exploited ones first, and filters over the CISA KEV catalog by vendor, product, ransomware use or date. Available in Ask AI, the CLI and MCP.
• MITRE ATT&CK mapping: curated CTID mappings for KEV-listed CVEs, plus techniques inferred from the CVE's weaknesses through CAPEC, labelled as inferred. Available to AI agents over MCP and through the REST API.
• IP, hash and domain intelligence from AbuseIPDB, GreyNoise, Shodan, VirusTotal, MalwareBazaar, crt.sh and more. Private and internal IPs are rejected before any lookup is made.
• Activity view: a searchable local history of your lookups, answers and the tool calls agents made, which you can pause or clear.
• Local Model view: install the runtime and download a model in one click.
• System view: health and version information.

Common uses

¹ For these questions the model has to pick the right lookup itself, which larger local models and cloud providers do more reliably. The same searches are always available from the terminal (see Recipes).

Install

Download the file for your OS from the latest release. The desktop app, CLI, MCP server and REST API are all one executable, with nothing else to install.

Each file has a .sha256 checksum next to it. Builds are not code-signed yet, so Windows SmartScreen and macOS Gatekeeper warn on first launch. On macOS, right-click the app and choose Open.

To build it yourself or run the headless server in Docker, see Build from source and Docker.

Local AI

Ask AI uses a small language model that runs on your computer, so your questions stay private. Before the model runs, CVE Intel finds the CVE IDs, IPs, hashes and domains in your question, runs the matching lookups and gives the model the results as facts to rely on. If no model is running, Ask returns those lookup results directly, so it is useful before any download finishes.

Open the Local Model view to install the llama.cpp runtime and download a model, one click each (files go under ~/.cve-intel/). The default is Qwen 2.5 3B Instruct, a ~1.9 GB download that needs 6 GB of RAM; the 1B and 1.5B models need 4 GB and Qwen 2.5 7B needs 10 GB. On Apple silicon the view also offers two MLX models that run on the GPU and answer faster. The model server listens on 127.0.0.1 only. Already running Ollama, LM Studio or your own llama-server? See Use your own model server.

Cloud providers (opt-in)

For stronger answers, choose OpenAI, Anthropic or Google Gemini under Settings → AI provider, pick a model and paste your API key. While one is selected, each question and its lookup results are sent to that vendor under their terms, and the Ask view shows a banner. Keys stay in ~/.cve-intel/settings.db with owner-only permissions and are never shown or logged (see SECURITY.md). All three are called through their OpenAI-compatible chat APIs; for Claude-specific features, use CVE Intel through MCP in a Claude client.

Data sources

The Databases view lists all 40 sources, and the CLI, REST API and MCP server report the same list.

Many tools need no API key at all: EPSS and its daily files, KEV, ENISA EUVD, CVE.org and CISA Vulnrichment, OSV, CVSS, Exploit-DB, Metasploit, PoC-in-GitHub, Project Zero, CISA advisories, CIRCL sightings, ATT&CK, CAPEC, crt.sh, Ransomwhere, hashlookup, Pwned Passwords, and NVD at a lower rate limit. Large feeds (EPSS daily files, Exploit-DB, Metasploit, ATT&CK, CAPEC, the CTID mappings) are downloaded once, reduced to a small index, and refreshed daily or every few hours.

Risk score

Risk = CVSS×0.20 + EPSS×0.35 + KEV×0.30 + PoC×0.15   (boosters up to ×1.27, capped at 100)

• KEV override: a CVE in CISA KEV is confirmed exploited, so it is always CRITICAL with a score of at least 76. Only CISA KEV earns the KEV points and the override.
• EPSS measures the probability of exploitation. CVSS is treated as severity, not likelihood.
• PoC is the strongest public exploit evidence. A Metasploit exploit module counts as weaponized (the full 15 points); Exploit-DB entries and VulnCheck XDB exploits count as public exploits.
• Deep triage also returns the SSVC v2 CISA Deployer decision: Act, Attend, Track* or Track. Its inputs come from CISA's published SSVC assessment when there is one. Because that assessment is a snapshot, newer evidence can raise Exploitation: a CISA KEV, VulnCheck KEV or ENISA EUVD exploited listing, or an EPSS of 0.90 or more, makes it active. EUVD and VulnCheck do not change the risk score.
• Bands: 0–25 LOW, 26–50 MEDIUM, 51–75 HIGH, 76–100 CRITICAL.
• This is scoring version 1.1, reported as scoring_version in results.

API keys

Keys are optional. Each one unlocks more data or raises a rate limit. In the app, open Settings → Data source API keys, press Get a key to open the sign-up page, paste the key and press Save. The key is used on the next lookup, with no restart needed.

Keys are stored in ~/.cve-intel/settings.db with owner-only permissions and shared by the app, CLI, MCP server and REST API on the same machine. A key saved in the app takes precedence over an environment variable. Free tiers and sign-up steps are set by each vendor, so check their pages. To manage keys from a terminal or with environment variables, see Terminal configuration.

Privacy and security

• All network traffic is outbound HTTPS: to the data sources, to the cloud AI provider if you opt in to one, and for the model and runtime downloads you start (Hugging Face, the llama.cpp release on GitHub and, for MLX, uv and a Python 3.12 build from GitHub plus mlx-lm from PyPI).
• API keys are never logged, cached, shown in the UI or returned by any tool.
• Private and reserved IPs are rejected before any network lookup.
• The local model server and the app's MCP server bind to loopback only.
• SQLite files are created on first use: cache.db, a disposable lookup cache; ~/.cve-intel/settings.db, which holds keys, provider settings and the watchlist auto-refresh interval; and ~/.cve-intel/watch.db, the local watchlist (items, snapshots and change events). Clearing the cache never touches settings or the watchlist.
• While the app is open, an optional background task can refresh the watchlist (every 6 hours by default). It only contacts the public sources needed for the items you watch, and you can turn it off in Settings. Nothing runs while the app is closed.
• The desktop app keeps an activity log in ~/.cve-intel/activity.db (owner-only). It records what you look up and ask, the answers, setting and model changes, the tools agents call through the app's MCP server, and which hosts each action contacted. API keys are never recorded, only key names. Upstream requests are stored as host and path, without query strings. Open Activity to search it, pause it or clear it. It keeps the latest 20,000 entries. The CLI and cve-intel mcp record nothing.

See SECURITY.md for the full security model and how to report a vulnerability.

Developer notes

Everything in this section is for using CVE Intel from a terminal, scripts, CI pipelines, other programs or AI agents, and for building it yourself. The desktop app needs none of it.

• The cve-intel executable: where it is and how one file runs every mode
• Command line (CLI): every command, with recipes for common tasks
• CI and automation: fail a build, upload SARIF to GitHub code scanning, JSON output
• REST API: endpoints and examples
• MCP server setup: connect AI agents by hand
• Terminal configuration: keys, environment variables, your own model server
• Build from source and Docker
• Architecture and Development and tests

The cve-intel executable

The desktop app and the command-line tool are the same file. Double-clicking it opens the app; running it with any argument starts the CLI, the MCP server or the REST server instead.

# macOS
"/Applications/CVE Intel.app/Contents/MacOS/cve-intel" triage CVE-2024-3400
# Linux
./CVE-Intel_<version>_linux-x64.AppImage triage CVE-2024-3400
# Windows (default per-user install)
"%LOCALAPPDATA%\CVE Intel\cve-intel.exe" triage CVE-2024-3400

On Windows the executable is a windowed program, so the terminal does not wait for it. To see the output in order, use start /wait "" "…\cve-intel.exe" triage CVE-2024-3400, or in PowerShell & "…\cve-intel.exe" … | Out-Host. MCP clients start it directly and are not affected.

For servers and containers there is also a headless build without the GUI (cargo build --release -p cve-server, which produces target/release/cve-intel). From a source checkout, replace cve-intel with cargo run -p cve-server --. The examples below assume cve-intel is on your PATH.

Command line (CLI)

cve-intel triage CVE-2024-3400 --depth deep
cve-intel ask "Should we patch CVE-2021-44228?" --model qwen2.5-3b
cve-intel quick-find 185.220.101.34
cve-intel scan deps .                          # every lockfile, manifest and SBOM here, ranked by risk
cve-intel scan deps . --fail-on kev            # exit 3 if anything is in CISA KEV
cve-intel fix lodash npm 4.17.15               # version that fixes every known vulnerability
cve-intel product confluence_data_center --vendor atlassian --version 8.5.0
cve-intel kev-search --vendor microsoft --ransomware-only --epss
cve-intel landscape --days 7                    # what's new; --markdown for a ticket
cve-intel watch add vendor fortinet
cve-intel watch refresh && cve-intel watch changes --unread
cve-intel decide record CVE-2025-12345 --verdict defer --until 2026-12-31 --auto --note "waiting for the vendor"
cve-intel decide list --state needs_attention   # decisions whose reasons stopped holding
cve-intel decide calibration                    # how your decisions held up
cve-intel compare CVE-2024-3400 CVE-2023-44487
cve-intel password-check                       # hidden prompt; only a 5-character hash prefix is sent to HIBP
cve-intel sources --probe                      # every data source, with live status

triage, fix, product, kev-search, landscape, watch, sources and ask print readable text; add --json for JSON. lookup, epss and kev print text, and the other lookup commands print JSON, ready for jq. Run cve-intel <command> --help for every option.

Recipes

CI and automation

Fail a CI build on the findings that matter. scan deps takes files and folders (searched like the app does), ranks the findings by risk and exits with code 3 when a gate trips:

cve-intel scan deps . --fail-on kev                 # anything in CISA KEV
cve-intel scan deps . --fail-on exploited           # in KEV, or a public or weaponized exploit
cve-intel scan deps . --fail-on high                # risk HIGH or CRITICAL
cve-intel scan deps . --fail-on-severity critical   # the advisory's own rating, ignoring risk

The risk score weighs exploitation, so a CVSS 9.8 that nobody exploits can rank LOW. Use --fail-on-severity to gate on advisory severity instead, or both. Without --strict, anything that could not be checked (a package OSV rejected, EPSS or the exploit indexes unreachable) is reported on stderr and does not fail the build. --enrich basic skips the exploit indexes (about 50 MB, downloaded once a day); on CI runners, cache CACHE_DB_PATH to avoid fetching them on every run.

--format markdown writes a summary table for $GITHUB_STEP_SUMMARY or a pull request comment, --format json the full result for jq, --format cyclonedx a CycloneDX 1.5 SBOM of everything scanned with each finding, its risk signals and your decisions attached (for Dependency-Track, a release artifact or an auditor; it works for scan image and scan host too), and -o FILE writes to a file:

cve-intel scan deps . --format markdown >> "$GITHUB_STEP_SUMMARY"
cve-intel scan deps requirements.txt --format json | jq '.summary'
cve-intel scan deps . --format cyclonedx -o sbom.cdx.json

Show the findings in GitHub code scanning (the Security tab and pull request annotations) by uploading SARIF:

# .github/workflows/dependencies.yml
on: [push, pull_request]
permissions:
  contents: read
  security-events: write
jobs:
  scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Scan the dependencies with CVE Intel
        run: cve-intel scan deps . --format sarif -o cve-intel.sarif
      - uses: github/codeql-action/upload-sarif@v3
        with:
          sarif_file: cve-intel.sarif
          category: cve-intel

This assumes cve-intel is on the runner's PATH (for example the release AppImage, or cargo build --release -p cve-server). Each vulnerability is a rule with an OSV link and a security-severity. Each result points at the file and line that declares the package, names the advisory and the fixed version, and carries the risk score, KEV status and EPSS. Findings in KEV or with HIGH or CRITICAL risk are errors whatever the advisory's own rating.

Get a machine-readable verdict for a ticket or dashboard:

cve-intel triage CVE-2024-3400 --depth deep --json | jq '.risk | {risk_score, risk_label, urgency}'

REST API

cve-intel serve starts an HTTP server on 127.0.0.1:8000 (set HOST and PORT to change it) with the REST API under /api/v1 and MCP at /mcp. The server the desktop app starts from MCP Server serves the same endpoints.

Errors return HTTP 400 with {"error": "…"}.

curl -s "http://127.0.0.1:8000/api/v1/triage/CVE-2024-3400?depth=deep" | jq '.risk'

curl -s -X POST http://127.0.0.1:8000/api/v1/tools/find_fixed_version \
  -H 'content-type: application/json' \
  -d '{"package": "lodash", "ecosystem": "npm", "version": "4.17.15"}'

curl -s -X POST http://127.0.0.1:8000/api/v1/tools/get_attack_mapping \
  -H 'content-type: application/json' -d '{"cve_id": "CVE-2024-34102"}'

• /api/v1/ask only uses the local model (or CVE_LLM_BASE_URL), never a cloud provider, so a caller cannot spend your cloud key.
• cve-intel serve has no authentication and allows cross-origin requests, so browser front ends can call it. That also means any web page open in your browser can reach it while it runs. Keep it on loopback, stop it when you're done, or put it behind your own authenticated reverse proxy. The server the desktop app starts refuses web pages instead (see below).

MCP server setup

MCP is an open protocol and works with many clients, not only Claude. The easiest setup is the MCP Server view in the desktop app (see MCP server for AI agents). To configure it by hand:

# HTTP: connect to the server started from the app, or from `cve-intel serve`
claude mcp add --transport http cve-intel http://127.0.0.1:8000/mcp

# stdio: the agent launches its own copy (macOS app path shown)
claude mcp add cve-intel -- "/Applications/CVE Intel.app/Contents/MacOS/cve-intel" mcp

# stdio, using the headless binary
claude mcp add cve-intel -- /absolute/path/to/target/release/cve-intel mcp

Ready-made configs for Claude Code, Claude Desktop, Cursor, VS Code and other clients are in examples/mcp/.

The server exposes 41 tools, resources (kev://catalog, epss://scores/{cve_id}, manifest://tool-hash) and prompts (patch_decision, compare_and_prioritize, dependency_triage). It never registers a sampling handler or issues sampling/createMessage. Older tool names that duplicated other tools (prioritize_cves, check_exploit_availability, scan_container_packages, scan_repo_secrets) are no longer listed but still work. The watchlist tools (list_watchlist, get_watchlist_changes) are read-only: agents cannot add, remove or refresh items. The scan history tools (list_scans, get_scan_findings) are read-only too, and only answer local callers: MCP over stdio and the server the app starts. cve-intel serve accepts requests from any web page, so it does not serve saved scans.

The server started from the app always binds to 127.0.0.1 and checks the Host, Origin and Sec-Fetch-Site headers to refuse cross-site browser requests, which protects against DNS rebinding. It has no sign-in, so any program on the same computer can use it with your saved keys. Stop it when you don't need it.

Terminal configuration

Manage API keys without the app:

cve-intel keys list                          # which keys are set and where from (never prints values)
cve-intel keys links                         # where to get each key
cve-intel keys set SHODAN_KEY                # hidden prompt, keeps the key out of shell history
echo "$KEY" | cve-intel keys set SHODAN_KEY  # or pipe it in from a script
cve-intel keys remove SHODAN_KEY

For headless use, copy .env.example to .env, or set environment variables. Everything works without keys, with fewer results. A key saved in the app takes precedence over an environment variable.

Use your own model server

Set CVE_LLM_BASE_URL to any OpenAI-compatible endpoint, such as Ollama (http://127.0.0.1:11434/v1), LM Studio or your own llama-server, and optionally CVE_LLM_MODEL to the model name. Ask then uses it instead of the built-in runtime. To use a llama-server or mlx_lm.server you installed yourself, set LLAMA_SERVER_BIN or MLX_LM_SERVER_BIN to its path. To use a cloud provider from the CLI: cve-intel ask "Should we patch CVE-2024-3400?" --provider anthropic.

Build from source and Docker

# Desktop app (debug)
cargo build -p cve-intel-desktop
./target/debug/cve-intel-desktop

# Installable bundle (.dmg / .exe / .AppImage)
cargo install tauri-cli --version "^2" --locked
cd apps/desktop/src-tauri && cargo tauri build

# Headless binary only (servers, Docker)
cargo build --release -p cve-server

Docker runs the headless MCP server over HTTP, plus the REST API:

docker build -t cve-intel .
docker run --rm -p 127.0.0.1:8000:8000 --env-file .env cve-intel

Architecture

crates/cve-core     Harness-agnostic engine: 38 source clients, risk scorer,
                    SSVC, CVSS, validators, lockfile/SBOM parsers, SARIF
                    and CycloneDX output, SQLite cache, settings DB, activity log and
                    watchlist, landscape briefing, source catalog,
                    llama.cpp and MLX manager, cloud-provider settings and
                    the Ask agent. No MCP, Tauri or CLI types.
crates/cve-server   The cve-intel binary: MCP (stdio and streamable HTTP),
                    REST (/api/v1) and the CLI, as thin adapters over cve-core.
apps/desktop        Tauri app linking cve-core and cve-server natively. The
                    same executable runs the GUI, the CLI and the MCP server,
                    and can host MCP over HTTP in-process. The model server
                    (llama-server, or mlx_lm.server on Apple silicon) is the
                    only child process.
examples/mcp        Client configs for common MCP clients.

Development and tests

cargo test -p cve-core -p cve-server   # scorer, SSVC, CVSS, validators, cache, parsers, agent routing, source catalog
cargo test -p cve-core -p cve-server -- --ignored   # live checks against the real sources (network)
cargo build --release                  # cve-intel and the desktop binary

See CONTRIBUTING.md for the full development setup and the checks CI runs.

Citing and attribution

Citing results in a report or ticket

CVE Intel gathers facts from public sources; it is not the source. When you record a decision, cite the primary source and the retrieval date, not the app or the AI's wording:

Example (illustrative values): CVE-2021-44228: CVSS 10.0 (NVD); listed in CISA KEV, added 2021-12-10 (CISA); EPSS 0.94 as of 2026-10-18 (FIRST). Retrieved 2026-10-18 with CVE Intel 0.3.0, scoring v1.1.

• Scores change. EPSS updates daily, KEV grows and NVD entries are revised. Results are cached for minutes to hours, and up to a day for some feeds. The Databases view shows each source's cache time.
• The risk score is CVE Intel's own. Cite it as "CVE Intel risk score (scoring v1.1)" alongside the underlying CVSS, EPSS and KEV values. cve-intel health and the System view show the version.
• An AI answer is not evidence. Local or cloud, a model can be wrong. Cite the data shown beside the answer and verify it against the source.

Citing this software

@software{cve_intel,
  title   = {CVE Intel: local security intelligence for CVE triage},
  author  = {{CVE Intel contributors}},
  year    = {2026},
  version = {0.3.0},
  url     = {https://github.com/Fazmin/cve-intel},
  license = {MIT OR Apache-2.0},
  note    = {Native Rust desktop app, CLI, MCP server and REST API}
}

Use the version you ran (cve-intel --version).

Data source terms

CVE Intel only queries these services. It does not own or republish their data, and their terms apply to anything you copy out of the app. Check each provider's current terms, especially before commercial use. This is not legal advice.

Local AI models are not part of this project and carry their own licences, such as the Llama 3.2 Community License and size-dependent Qwen licences. Read the model's page before commercial use. GGUF files come from the Hugging Face repositories listed in crates/cve-core/src/llm.rs (the vendors' own for Qwen, bartowski's builds for Llama), and MLX builds from the mlx-community conversions. The runtimes are llama.cpp (MIT) and, on Apple silicon, mlx-lm (MIT).

Community

• CONTRIBUTING.md: development setup and pull request guidelines
• CODE_OF_CONDUCT.md: community standards
• SECURITY.md: reporting a vulnerability privately

License

Copyright 2024-2025 Fazmin Nizam. Licensed under either of

• Apache License, Version 2.0 (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
• MIT License (LICENSE-MIT or https://opensource.org/licenses/MIT)

at your option. See also NOTICE.

These licenses cover the CVE Intel code. Data and models the app downloads keep their own terms; see Data source terms.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.