Engineer working on a laptop connected to a device test rig, surrounded by wiring harnesses and test hardware

Firmware vulnerabilities & fixes

MITRE CAPEC: Common Attack Pattern Enumeration and Classification, explained

Last updated:

CAPEC (Common Attack Pattern Enumeration and Classification), currently List Version 3.9 (released 24 January 2023) with 559 total attack patterns, is MITRE's catalog of how adversaries actually carry out an attack, not just which weakness they exploit or which specific vulnerability they target. Per MITRE's own About page, it is "a publicly available catalog of common attack patterns that helps users understand how adversaries exploit weaknesses in applications and other cyber-enabled capabilities." Originally established by the US Department of Homeland Security as part of its Software Assurance strategic initiative, CAPEC is now sponsored by DHS's Cybersecurity and Infrastructure Security Agency (CISA) and managed by the Homeland Security Systems Engineering and Development Institute (HSSEDI), which MITRE operates — the same sponsorship structure that funds CWE and CVE.

Key takeaways

  • Every CAPEC entry follows the same structure, defined by the catalog's XML schema.
  • Domains of Attack — groups attack patterns "by the target domains" they apply to: Software, Hardware, Communications, Supply Chain, Social Engineering and Physical Security.
  • The link runs in both directions.
  • CAPEC is deliberately scoped differently from MITRE's other well-known framework, ATT&CK.

What does an "attack pattern" actually document?

MITRE's "New to CAPEC" page is explicit about where CAPEC sits relative to the other weakness/vulnerability catalogs a firmware or application security team already tracks: "a CAPEC attack pattern is typically a method of leveraging a CWE to execute an attack." Each entry documents "step-by-step instructions for an adversary to explore for potential targets, experiment with their assets and defensive mechanisms, if any, and then to carry out the exploit." The important distinction MITRE draws is that CAPEC entries describe method, not outcome: "CAPEC entries are not based on the consequences of an attack, but on how to exploit a weakness to cause the consequence." In practice, that gives a security team three distinct, cross-referenced layers to work with:

Catalog What it names
CWE The underlying weakness class (e.g. improper input validation) — see our CWE Top 25 guide
CAPEC The reusable adversary method for turning that weakness into a working attack
CVE One specific, concrete instance of a vulnerability in a named product/version

What is inside a single CAPEC entry?

Every CAPEC entry follows the same structure, defined by the catalog's XML schema. Taking CAPEC-88 OS Command Injection as the reference, the fields a firmware team uses most are:

Field What it holds (CAPEC-88 as example) Use for a device team
Description An adversary "injects operating system commands into existing application functions" Scoping: is this method possible on our product at all?
Likelihood Of Attack / Typical Severity High / High A first, generic signal; your own risk method still decides
Execution Flow Four steps across the Explore, Experiment and Exploit phases, from "Identify inputs for OS commands" to "Execute malicious commands" Test-case skeleton for a penetration test
Prerequisites "User controllable input used as part of commands to the underlying operating system" Code-review question: where does input reach a shell?
Skills Required Rated High for CAPEC-88 Who could realistically do this
Consequences Execute unauthorized commands, gain privileges, read data Impact statement in a threat model
Mitigations Use language APIs instead of the shell, filter input, run with least privilege Design requirements and fix guidance
Related Weaknesses CWE-78, CWE-88, CWE-20, CWE-697 The bridge back to code findings
Taxonomy Mappings WASC, OWASP, and for many entries ATT&CK The bridge to detection and threat intel

Not every entry is complete. Many are marked Draft, and some lack an execution flow or a likelihood rating — CAPEC-70, for example, has description, prerequisites, skills and mitigations but no step-by-step flow. Read the entry status before treating it as a finished test plan.

How is the catalog organized?

The current catalog is CAPEC List Version 3.9, released 24 January 2023 and, as of this guide's publication, still the live version — MITRE's list page and the downloadable XML both show 3.9 when checked on 12 November 2026. It contains 559 total attack patterns; the XML file holds 615 attack-pattern IDs, of which 56 are marked Deprecated. The active 559 split into 61 Meta, 181 Standard and 317 Detailed entries. They are browsable through two independent classification views plus a three-tier abstraction scale:

  • Mechanisms of Attack — MITRE's own description: this view "organizes attack patterns hierarchically based on mechanisms that are frequently employed when exploiting a vulnerability." Its nine top-level categories are Engage in Deceptive Interactions, Abuse Existing Functionality, Manipulate Data Structures, Manipulate System Resources, Inject Unexpected Items, Employ Probabilistic Techniques, Manipulate Timing and State, Collect and Analyze Information, and Subvert Access Control.
  • Domains of Attack — groups attack patterns "by the target domains" they apply to: Software, Hardware, Communications, Supply Chain, Social Engineering and Physical Security. A device maker needs all six, unlike a pure web team that lives mostly in Software.
  • Abstraction levels — every entry is tagged Meta, Standard, or Detailed, running from broad, high-level attack concepts down to narrow, highly specific, step-by-step patterns; a Meta pattern groups several related Standard or Detailed patterns underneath it.

CAPEC also publishes focused views, including a Supply Chain Risks view (CAPEC-683) and the ICS view covered below.

The link runs in both directions. In the v3.9 XML, 450 of the 559 active entries list at least one Related Weakness, and each CWE page carries a "Related Attack Patterns" section pointing back. The CWE-78 page, for instance, lists CAPEC-108, CAPEC-15, CAPEC-43, CAPEC-6 and CAPEC-88. A practical walk from a vulnerability to attack patterns looks like this:

  1. Start from the CVE record and note its CWE ID (the CVE or NVD record usually carries one).
  2. Open that CWE page on cwe.mitre.org and read the Related Attack Patterns list.
  3. Open each CAPEC entry, check its abstraction level and status, and read Prerequisites against your product.
  4. Keep the entries whose prerequisites hold; use their Execution Flow as test steps and their Mitigations as fix criteria.
  5. Follow any ATT&CK Taxonomy Mapping on those entries to see how the method shows up in real intrusions.

The gaps matter as much as the links. About a fifth of active entries list no CWE at all — many hardware and supply-chain patterns, such as CAPEC-638 Altered Component Firmware, describe methods that do not reduce to a single code weakness.

How does CAPEC differ from ATT&CK?

CAPEC is deliberately scoped differently from MITRE's other well-known framework, ATT&CK. MITRE's own published comparison draws the line at the application-security/network-defense boundary:

CAPEC ATT&CK
Scope "focused on application security and describes the common attributes and techniques employed by adversaries to exploit known weaknesses" "focused on network defense and describes the operational phases in an adversary's lifecycle"
Primary use cases Application threat modeling, developer training and education, penetration testing Comparing network defense capabilities, defending against advanced persistent threats, threat hunting

MITRE notes the two are complementary rather than overlapping: "many attack patterns enumerated by CAPEC are employed by adversaries through specific techniques described by ATT&CK," and the catalogs carry cross-references so a CAPEC entry can point to the ATT&CK technique(s) an adversary would use to execute it operationally. In v3.9, 177 active entries carry an ATT&CK mapping. CAPEC-70, for example, maps to ATT&CK T1078.001 Default Accounts. For the defensive side of the same picture, see our D3FEND guide.

Which CAPEC patterns matter for embedded devices?

The table below is a starting set for connected devices and firmware, taken from the v3.9 XML on 12 November 2026. It is a selection, not a complete list.

CAPEC entry Abstraction Related CWE ATT&CK mapping
CAPEC-70 Try Common or Default Usernames and Passwords Detailed CWE-521, CWE-798 and others T1078.001
CAPEC-49 Password Brute Forcing Standard CWE-521, CWE-307 and others T1110.001
CAPEC-88 OS Command Injection Standard CWE-78, CWE-88, CWE-20, CWE-697 —
CAPEC-126 Path Traversal Standard CWE-22 —
CAPEC-100 Overflow Buffers Standard CWE-120, CWE-119 and others —
CAPEC-186 Malicious Software Update Standard CWE-494 T1195.002
CAPEC-657 Malicious Automated Software Update via Spoofing Detailed CWE-494 T1072
CAPEC-638 Altered Component Firmware Detailed — T1542.002
CAPEC-532 Altered Installed BIOS Detailed — T1495, T1542.001
CAPEC-624 Hardware Fault Injection Meta CWE-1247, CWE-1332 and others —
CAPEC-702 Exploiting Incorrect Chaining or Granularity of Hardware Debug Components Detailed CWE-1296 —
CAPEC-94 Adversary in the Middle (AiTM) Meta CWE-300, CWE-287 and others T1557

Our fix guides cover the code side of several rows: command injection, and OTA update security for the CAPEC-186 / CAPEC-657 pair.

What does CAPEC look like on a real firmware finding?

Worked example — synthetic, for illustration only. A router team's firmware SBOM match shows a bundled web-admin component with a published CVE whose record lists CWE-78. The team wants to know whether the issue is exploitable on their build and what to test.

  1. CWE to CAPECCWE-78's Related Attack Patterns give five entries. CAPEC-108 (SQL injection to command line) does not apply: the device has no SQL database. CAPEC-88 and CAPEC-15 Command Delimiters do.
  2. PrerequisitesCAPEC-88 needs "user controllable input used as part of commands." The team finds that the diagnostics page passes a hostname field to ping through a shell. The prerequisite holds.
  3. Execution flow as a test planThe four CAPEC-88 steps become four test cases: list every input that reaches a shell; survey the admin UI as an authenticated user; send delimiter payloads (from CAPEC-15) in each field; confirm whether a command runs.
  4. Mitigations as fix criteriaThe fix ticket quotes CAPEC-88's mitigations: call the binary through an exec-style API with an argument array, validate the hostname, and drop the web server's root privileges.
  5. PriorityCAPEC's High likelihood and severity are generic. The team sets priority from its own risk method and from exploit data — whether the CVE is in CISA KEV and its EPSS score; see our KEV and EPSS guide.

The outcome is a scoped, repeatable test and a fix definition, both traceable to public catalog entries.

Is there a CAPEC view for ICS and OT?

Yes. For teams working on operational technology, CAPEC publishes CAPEC-703, "Industrial Control System (ICS) Patterns," a curated view whose stated objective is "a listing of CAPECs that apply to industrial control systems (ICS)," noting that "some children of these attack patterns might also be applicable" even where not explicitly listed. As of CAPEC v3.9, this view groups 46 of the catalog's 559 attack patterns, built in part using the ATT&CK for ICS matrix as an input, and spans the same Meta-to-Detailed abstraction range as the full catalog — from broad patterns like adversary-in-the-middle and flooding down to detailed patterns tied to specific ICS-relevant exploitation techniques. It gives an OT security team a pre-filtered starting point instead of manually screening all 559 entries for industrial relevance.

Can CAPEC content be reused commercially?

MITRE's Terms of Use for CAPEC state: "The MITRE Corporation (MITRE) hereby grants you a non-exclusive, royalty-free license to use Common Attack Pattern Enumeration and Classification (CAPEC™) for research, development, and commercial purposes," conditioned on reproducing "MITRE's copyright designation and this license in any such copy." That means CAPEC content can be reused, including commercially, as long as attribution and the licence text travel with the copy — the same free-to-use-with-attribution model MITRE applies to CWE and ATT&CK. The full catalog is on the downloads page as XML and CSV, which is the easiest way to load it into a threat-modelling tool or test-management system.

What are the common pitfalls?

  • Treating Likelihood and Severity as your risk score. They are generic ratings for the method, not for your product.
  • Citing Meta patterns as test cases. Meta entries such as CAPEC-624 describe a family; test against the Standard or Detailed children.
  • Stopping at CWE-linked entries. Hardware, physical and supply-chain patterns often have no CWE; a device threat model needs them anyway.
  • Ignoring status. Draft entries can lack execution flows or mitigations; fill the gap from your own engineering, not from the entry name.
  • Assuming the catalog moves fast. v3.9 dates from January 2023; newer attack methods may only appear in ATT&CK or EMB3D.

Where should a device team start?

  1. Use CAPEC to fill the gap between "what's wrong" and "what an attacker actually does."If your team already tracks CWE classes and CVE instances, CAPEC is the missing middle layer that documents the reusable method connecting the two — valuable input for threat modeling sessions and penetration-test scoping. Our threat modelling guide shows where it fits in a session.
  2. Start from CAPEC-703 if your focus is OT/ICS, rather than screening the full 559-entry catalog — it's a pre-scoped 46-pattern starting set built with ATT&CK for ICS as an input.
  3. Cross-reference CAPEC entries against your own CWE/CVE tracking, the same discipline our EMB3D and CWE Top 25 guides recommend for MITRE's other knowledge bases — CAPEC is explicitly designed to interlock with both.
  4. Treat CAPEC as a penetration-testing and threat-modeling input, not an incident-response taxonomy— for the adversary-lifecycle/TTP view used in detection engineering and threat hunting, MITRE's own guidance points to ATT&CK instead.
  5. Build the device table onceKeep a short list like the one above, filtered by the Hardware, Communications and Supply Chain domains, and reuse it per product line. Our firmware security testing pillar shows how to turn it into a test plan.
  6. Check the list version before citing a specific CAPEC-ID in a documentCAPEC v3.9 (24 January 2023) is current as of this guide, but MITRE revises the catalog periodically; verify against capec.mitre.org directly before relying on an entry's exact wording or ID in an audit artifact.

SoftWSP covers the step before CAPEC: it extracts an SBOM from a firmware image, matches the components against published CVEs in NVD, CISA KEV, FIRST EPSS and CISA CSAF ICS advisories, and drafts a cited report for an engineer to review. The CWE behind each CVE comes from the public record; choosing the CAPEC patterns and test cases that follow is your team's call, or part of a scoped engagement: see device & firmware security.

This guide summarises MITRE's own published CAPEC web site, the CAPEC v3.9 XML and Terms of Use; it is not a substitute for reading the current capec.mitre.org content directly, and does not constitute security or compliance advice.

Put this guide to work

Frequently asked questions

What is CAPEC?

CAPEC (Common Attack Pattern Enumeration and Classification) is a free, publicly available catalog of common attack patterns. MITRE's own About page describes it as helping "users understand how adversaries exploit weaknesses in applications and other cyber-enabled capabilities," documenting "the common attributes and approaches employed by adversaries to exploit known weaknesses." It was established by the US Department of Homeland Security as part of the Software Assurance strategic initiative, and per MITRE's own sponsorship model (the same structure used for CWE and CVE) it is currently sponsored by DHS's Cybersecurity and Infrastructure Security Agency (CISA) and managed by the Homeland Security Systems Engineering and Development Institute (HSSEDI), which MITRE operates.

What exactly is an "attack pattern", and how is it different from a CWE or a CVE?

Per MITRE's own "New to CAPEC" page, "a CAPEC attack pattern is typically a method of leveraging a CWE to execute an attack," with "step-by-step instructions for an adversary to explore for potential targets, experiment with their assets and defensive mechanisms... and then to carry out the exploit." MITRE is explicit that "CAPEC entries are not based on the consequences of an attack, but on how to exploit a weakness to cause the consequence" — a CWE (see our [CWE Top 25 guide](/guides/cwe-top-25-software-weaknesses/)) names the underlying flaw class, a CVE names one specific instance of a vulnerability, and a CAPEC attack pattern documents the reusable adversary *method* that turns a given CWE into a working exploit.

How is CAPEC structured, and how big is the current list?

The current list is CAPEC List Version 3.9, released 24 January 2023 and still the live version as of this guide, containing 559 total attack patterns. Patterns are organized along two independent views — Mechanisms of Attack, which "organizes attack patterns hierarchically based on mechanisms that are frequently employed when exploiting a vulnerability," and Domains of Attack, which groups patterns "by the target domains" they apply to — and across three abstraction levels, Meta, Standard and Detailed Attack Patterns, from broad/high-level patterns down to highly specific, step-by-step ones.

How does CAPEC relate to MITRE ATT&CK?

They are complementary, not duplicates. MITRE's own comparison states CAPEC is "focused on application security and describes the common attributes and techniques employed by adversaries to exploit known weaknesses," useful for "application threat modeling, developer training and education, [and] penetration testing," while [ATT&CK](/guides/mitre-attack-enterprise/) is "focused on network defense and describes the operational phases in an adversary's lifecycle," useful for "comparing computer network defense capabilities" and "hunting for new threats." MITRE adds that "many attack patterns enumerated by CAPEC are employed by adversaries through specific techniques described by ATT&CK," and the two catalogs carry cross-references so a given CAPEC entry can point to the ATT&CK technique(s) that realize it in the field.

Is there a CAPEC view specific to industrial control systems / OT?

Yes. CAPEC-703, "Industrial Control System (ICS) Patterns," is a dedicated view whose stated objective is to contain "a listing of CAPECs that apply to industrial control systems (ICS)" (noting "some children of these attack patterns might also be applicable"). As of CAPEC v3.9 it groups 46 of the catalog's 559 attack patterns, built in part using the [ATT&CK for ICS](/guides/mitre-attack-for-ics/) matrix, and spans abstraction levels from meta-level patterns (e.g. adversary-in-the-middle, flooding) down to detailed patterns addressing specific ICS-relevant techniques.

Sources

  1. MITRE CAPEC — About CAPEC (sponsorship, HSSEDI management)
  2. MITRE CAPEC — New to CAPEC (attack pattern definition, relation to CWE/CVE)
  3. MITRE CAPEC — CAPEC and ATT&CK (official comparison)
  4. MITRE CAPEC — CAPEC List Version 3.9 (current list landing page, 559 patterns)
  5. MITRE CAPEC — CAPEC-703: Industrial Control System (ICS) Patterns view
  6. MITRE CAPEC — About CAPEC: Terms of Use (non-exclusive, royalty-free licence)
  7. CVE.org — official CVE Program site
  8. MITRE CAPEC — Downloads (XML and CSV of the full catalog)
  9. MITRE CAPEC — capec_latest.xml (List Version 3.9, 2023-01-24)
  10. MITRE CAPEC — CAPEC-1000: Mechanisms of Attack view
  11. MITRE CAPEC — CAPEC-3000: Domains of Attack view
  12. MITRE CAPEC — CAPEC-683: Supply Chain Risks view
  13. MITRE CAPEC — CAPEC-88: OS Command Injection
  14. MITRE CAPEC — CAPEC-70: Try Common or Default Usernames and Passwords
  15. MITRE CAPEC — CAPEC-186: Malicious Software Update
  16. MITRE CAPEC — CAPEC-638: Altered Component Firmware
  17. MITRE CAPEC — CAPEC-624: Hardware Fault Injection
  18. MITRE CWE — CWE-78: OS Command Injection (Related Attack Patterns)
  19. MITRE CWE — CWE-494: Download of Code Without Integrity Check
  20. MITRE ATT&CK — T1078.001 Valid Accounts: Default Accounts
Was this guide helpful?

In this cluster

MITRE CAPEC: Common Attack Pattern Enumeration and Classification, explained