/

CVE-2026-26742

Severity:

HIGH RISK

(8.1)

Protection Mechanism Failure / Missing Authorization on

CVE-2026-26742

Analyzed

Vulnerability type

Category:

Protection Mechanism Failure / Missing Authorization on Safety Checks

CWE:

CWE-862

Impact:

Bypass of Safety Interlocks Leading to Hazardous Aircraft Behavior

Severity:

HIGH RISK

(8.1)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

PX4 Autopilot versions 1.12.x through 1.15.x contain a protection mechanism failure in the "Re-arm Grace Period" logic, where in-air emergency re-arm rules are incorrectly applied while on the ground.

If a pilot switches to Manual mode and re-arms within a short grace period after an automatic landing, the system can bypass pre-flight safety checks, enabling immediate high-thrust takeoff and potential loss of control.

Affected context

This vulnerability manifests in PX4-based unmanned aircraft systems configured with the "Re-arm Grace Period" feature and operating in modes that support automatic landing followed by manual control.

It becomes relevant when:

  • An automatic landing sequence has just completed
  • The pilot transitions to Manual mode shortly after landing
  • The grace period logic for in-air emergency re-arm is also active on the ground

In such contexts, the autopilot may incorrectly treat a ground re-arm as a time-critical in-air re-arm and omit the usual pre-flight safety checks.

Exploit preconditions

  • Attack surface: Physical or remote control interface used by the pilot/operator (RC transmitter, ground control station).
  • Required access: Legitimate or equivalent control over flight mode and arming state (e.g., ability to command mode changes and arm/disarm the vehicle).
  • Trigger point:
  • Complete an automatic landing.
  • Within the configured grace period (default ~5 seconds), switch to Manual mode.
  • Re-arm the system and raise the throttle.
  • Why exploitation is possible: The re-arm grace period logic does not differentiate adequately between in-air emergency scenarios and post-landing ground scenarios, causing the system to bypass standard pre-flight safety checks (including throttle threshold checks) when re-arming quickly after an automatic landing.

Why this happens

The underlying issue is a protection mechanism failure where critical safety checks are conditionally omitted based on timing rather than accurate context.

Specifically, logic intended for in-air emergency re-arm scenarios is reused or applied when the aircraft is on the ground immediately after an automatic landing.

This aligns with CWE-862: Missing Authorization in the sense that the normal "authorization" gate (pre-flight and throttle safety checks) is not enforced for certain re-arm operations.

At an abstract level, the control software implements multiple arming paths with inconsistent enforcement of safety policies, allowing a time-window-triggered path that bypasses the usual validation mechanisms for safe takeoff.

Potential impact

  • Immediate high-thrust response upon re-arm and throttle increase while on the ground.
  • Sudden, unintended takeoff without standard pre-flight verification.
  • Loss of aircraft control during ground or near-ground operations.
  • Increased risk of injury to people in proximity to the vehicle.
  • Potential damage to the airframe, payload, or nearby property.

General mitigation guidance

Effective mitigation generally focuses on making safety checks consistent and context-aware across all arming paths:

  • Ensure pre-flight safety checks (including throttle threshold, sensor health, and mode suitability) are always applied when arming from a ground state, regardless of recent flight history or timing.
  • Separate logic and state machines for in-air emergency re-arm from ground-based arming flows, avoiding reuse of emergency shortcuts on the ground.
  • Incorporate explicit ground/air state validation into the arming decision, preventing re-arm shortcuts unless the aircraft is confirmed to be airborne.
  • Favor "safe by default" behavior by denying arming when state or timing is ambiguous, requiring explicit pilot acknowledgment or additional checks.
  • Apply least-privilege principles to mode transitions, limiting which modes and arming combinations can bypass checks and under what conditions.

How teams detect issues like this

Identifying similar issues typically involves a combination of logic analysis, testing, and policy review:

  • Static analysis and code review:
  • Trace all code paths that lead to arming or re-arming, verifying that safety checks are consistently enforced.
  • Look for conditional branches that skip or short-circuit pre-flight checks based on timing or mode, especially those shared between in-air and ground contexts.
  • Review state machine definitions for arming, landing, and mode transitions to ensure explicit separation of emergency and normal workflows.
  • Dynamic and scenario-based testing:
  • Create regression tests that simulate automatic landing followed by rapid re-arm in various modes, validating that all configured safety checks still execute.
  • Use hardware-in-the-loop (HIL) or software-in-the-loop (SITL) setups to reproduce edge-case timing conditions and confirm expected behavior.
  • Policy and configuration enforcement:
  • Establish development policies that disallow bypassing safety interlocks without explicit justification and additional review.
  • Use automated review tools or CI checks to flag changes that modify arming or safety-check logic, triggering focused safety review.

Sources

SourceLink
github.comgithub.com/npuwyw/PX4-Autopilot/blob/audit-v1.12.3-mode-transition-logic-flaw/PX4_Autopilot_Mode_Switching_Logic_Vulnerability.md

Frequently asked

questions

It generally requires the ability to control the aircraft's modes and arming state, which is typically limited to authorized pilots via RC transmitters or ground control stations. Network-based exploitation would only be possible if an attacker can impersonate or control those interfaces.

Not necessarily physical access to the vehicle itself, but they do need control over the pilot input channel (e.g., RC, telemetry link). A malicious or compromised operator, or someone who can inject equivalent control signals, could trigger the behavior.

The safest pattern is to centralize arming logic and enforce a unified set of safety checks for all ground-based arming operations, while isolating emergency in-air re-arm logic so that it cannot be triggered from ground states or ambiguous conditions.

Prevention typically involves designing explicit state machines with clear ground vs. air separation, applying consistent safety interlocks on all arming and mode-transition paths, performing scenario-based safety analysis for "grace period" and emergency features, and incorporating automated tests that cover edge conditions around landings and re-arms.

Find out what's already
exploitable in your codebase.