Eviulon-controlled external publication · source-linked · no behavioral analytics InternationalIntelligence.org

EVIDENCE STATES / PUBLIC UNKNOWNS / DEGRADED MODES

Kill Webs Evidence and Degraded-Mode Observatory

Trace forty-seven evidence-qualified findings to the thirteen retained reports, then test how a synthetic mission thread should change behavior when identity, time, semantics, transport, provenance, or assessment can no longer be trusted.

47 qualified findings 13 retained reports 118 finding-to-report connections Reviewed: Aug 3, 2026

DIRECT ANSWER / EVIDENCE BOUNDARY

Observatory in brief

Core distinction
The chain is the selected mission process; the web is the changing option space. Neither is resilient unless evidence, authority, and dependencies remain trustworthy.
Public record
47 findings are mapped to 13 retained reports with explicit evidence-state labels.
Degraded-mode rule
A function that remains online may still be too stale, disputed, unauthenticated, or semantically uncertain to support the next consequential transition.

REGISTER SUMMARY

A claim ledger across the complete Kill Webs collection

The register is derived from the evidence-qualified findings already published on the eight research routes. It does not create a second editorial record or silently upgrade recommendations into facts.

Qualified findings
47
Research routes
8
Retained reports
13
Evidence connections
118
Publicly unspecified findings
3

TRACEABLE FINDINGS

Filter by evidence state, route, or retained report

Filters change only this view. Each finding remains governed by the wording, cautions, and scope of its full research route.

Showing 47 findings

SUPPORTED BY SUBMITTED REPORTS18

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

CROSS-REPORT ANALYSIS4

This is a cross-report synthesis. It is not a direct quotation, a universal doctrinal definition, or proof that every implementation behaves this way.

EVIDENCE CAUTION6

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

RECOMMENDED CONTROL16

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

PUBLICLY UNSPECIFIED3

The submitted public material does not establish this detail. Unknown must not be silently converted into either a reassuring or alarming claim.

01

SUPPORTED BY SUBMITTED REPORTS

A web composes chains; it does not abolish them

Concepts and orchestration

The submitted reports consistently retain the ordered find, fix, track, target, engage, and assess functions; the architectural change is that those functions need not be bound in advance to one sensor, headquarters, network, and effector. [13] [1] [8]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
3
Open the full research route
02

CROSS-REPORT ANALYSIS

A web composes chains; it does not abolish them

Concepts and orchestration

At the moment of execution, one pathway through the web still behaves as a chain. The web changes when and how that pathway is assembled, compared, rerouted, and audited. [13] [12]

How to read this state

This is a cross-report synthesis. It is not a direct quotation, a universal doctrinal definition, or proof that every implementation behaves this way.

Retained reports
2
Open the full research route
03

EVIDENCE CAUTION

A web composes chains; it does not abolish them

Concepts and orchestration

A physically distributed architecture can remain logically centralized if every route depends on one cloud, identity service, fusion engine, gateway, classification domain, or approval authority. [13] [7]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
2
Open the full research route
04

SUPPORTED BY SUBMITTED REPORTS

A web composes chains; it does not abolish them

Concepts and orchestration

The recurring strategic claim is optionality under disruption: losing a preferred node should narrow the option set rather than automatically terminate the mission. [1] [8] [12]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
3
Open the full research route
05

A web composes chains; it does not abolish them

Concepts and orchestration

The practical objective should be the best authorized, compatible, sufficiently trusted, and operationally available pathway—not a literal rule that every sensor must connect to every shooter. [13] [12]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
2
Open the full research route
06

SUPPORTED BY SUBMITTED REPORTS

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

The reports describe ACK as adapting commercial marketplace, sourcing, and supply-chain ideas: mission demand appears as a consumer need, while organizations advertise available capabilities as suppliers. [1] [8]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
07

SUPPORTED BY SUBMITTED REPORTS

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

The Virtual Liaison abstraction is presented as a way to expose an effect, availability, and constraints without necessarily disclosing every sensitive source, method, or internal implementation detail. [1] [8]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
08

PUBLICLY UNSPECIFIED

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

The reports say ACK reduced some reallocation and assignment decisions to the order of minutes and transitioned technology to military Services, while the receiving programs, fielded configurations, and operational extent remain incompletely public. [1] [8]

How to read this state

The submitted public material does not establish this detail. Unknown must not be silently converted into either a reassuring or alarming claim.

Retained reports
2
Open the full research route
09

CROSS-REPORT ANALYSIS

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

ACK and integration middleware solve different problems: orchestration selects a candidate combination; translation and gateway technologies make heterogeneous systems exchange usable data. [1] [7] [12]

How to read this state

This is a cross-report synthesis. It is not a direct quotation, a universal doctrinal definition, or proof that every implementation behaves this way.

Retained reports
3
Open the full research route
10

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

Any marketplace recommendation must remain subordinate to identity, data quality, capacity, release, command authority, rules, safety, and consequence constraints enforced outside the optimization score. [1] [11]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
2
Open the full research route
11

Discover, compare, and reallocate capabilities across boundaries

Concepts and orchestration

A defensible marketplace evaluates hard eligibility before comparative utility: authority, identity, provenance, releasability, and assessment should be fail-closed gates, while timeliness, resilience, capacity, auditability, and reversibility may inform ranking among the remaining options. [1] [11] [13]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
12

SUPPORTED BY SUBMITTED REPORTS

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

The reports consistently frame JADC2 or CJADC2 as broader than targeting: it includes data, transport, analytics, command structures, policy, authorities, organizations, training, partners, and operational procedures. [12] [6] [1]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
3
Open the full research route
13

SUPPORTED BY SUBMITTED REPORTS

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

Service-specific initiatives contribute different sensing, transport, command, fires, maritime, expeditionary, and space capabilities, but joint value depends on agreed interfaces, semantics, identity, and operating procedures. [12] [6]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
14

EVIDENCE CAUTION

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

A common operating picture should provide role-tailored views over consistent underlying data; identical screens for every user can be as misleading as incompatible pictures. [12]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
1
Open the full research route
15

SUPPORTED BY SUBMITTED REPORTS

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

Space-based warning, tracking, transport, communications, and positioning are described as critical enablers, while public reports also retain integration, schedule, ground-processing, and demonstration uncertainties. [12] [6]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
16

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

Coalition integration requires explicit classification, releasability, provenance, identity, and policy controls; a technically reachable partner node is not automatically an authorized participant in every mission thread. [12] [6] [13]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
17

JADC2 is the enterprise; the kill web is one composable operational thread

Joint enterprise and technical backbone

Role-tailored common operating pictures should preserve one shared record identity, source time, uncertainty, authority, and assessment state even when each partner receives a different level of technical detail. [12] [6] [13]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
18

SUPPORTED BY SUBMITTED REPORTS

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

The reports converge on a machine-readable capability model: a node must describe what it can provide, under which conditions, at what quality, with which dependencies, and under whose authority. [1] [7] [12]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
3
Open the full research route
19

SUPPORTED BY SUBMITTED REPORTS

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

A federated data fabric must carry time, position, uncertainty, source provenance, classification, releasability, confidence, and update age—not merely a coordinate or label. [1] [12]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
20

SUPPORTED BY SUBMITTED REPORTS

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

Resilient transport and edge computing reduce dependence on central clouds by prioritizing compact, mission-relevant updates and local continuity when bandwidth or long-haul links degrade. [7] [12] [6]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
3
Open the full research route
21

PUBLICLY UNSPECIFIED

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

The reports reject one universal kill-web latency threshold. Local control, tactical cueing, human decision support, and cross-domain resource allocation operate on different time scales and require separate budgets. [1] [8]

How to read this state

The submitted public material does not establish this detail. Unknown must not be silently converted into either a reassuring or alarming claim.

Retained reports
2
Open the full research route
22

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

Open interfaces and translation middleware should be paired with conformance tests, data-rights strategy, identity controls, supply-chain assurance, and mission-thread exercises under simultaneous degradation. [7] [1] [10]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
23

Semantics, transport, edge computing, identity, interfaces, and test

Joint enterprise and technical backbone

Interface conformance must test field meaning, units, time basis, uncertainty, provenance, authority, version behavior, and declared information loss—not only whether a message parses. [7] [12] [9]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
24

SUPPORTED BY SUBMITTED REPORTS

A connected web can fail silently through corrupted trust and data

Resilience, integrity, and autonomy

Greater interoperability expands the number of interfaces, credentials, gateways, translators, update paths, software dependencies, and trust relationships that must be defended. [10] [3]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
25

SUPPORTED BY SUBMITTED REPORTS

A connected web can fail silently through corrupted trust and data

Resilience, integrity, and autonomy

Integrity attacks can be more consequential than visible denial because plausible false tracks, manipulated confidence, stale telemetry, or altered timing can induce confident but incorrect decisions. [10] [3]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
26

SUPPORTED BY SUBMITTED REPORTS

A connected web can fail silently through corrupted trust and data

Resilience, integrity, and autonomy

Position, navigation, and time are hidden shared dependencies: small or gradual errors can disrupt association, fusion, sequencing, authentication, and confidence across many nodes. [10] [3]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
27

EVIDENCE CAUTION

A connected web can fail silently through corrupted trust and data

Resilience, integrity, and autonomy

Integration gateways and translation layers solve interoperability problems while creating new places where meaning, identity, version, or policy can be lost or manipulated. [3] [7]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
2
Open the full research route
28

A connected web can fail silently through corrupted trust and data

Resilience, integrity, and autonomy

The defensible objective is mission resilience rather than perfect prevention: assume some nodes, links, credentials, datasets, and models will become unavailable or untrustworthy, then constrain and rehearse the response. [10] [3]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
2
Open the full research route
29

SUPPORTED BY SUBMITTED REPORTS

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

The reports describe a layered AI pipeline: sensor-specific perception feeds tracking and fusion; those outputs feed prediction and constrained resource assignment. Each layer adds uncertainty and potential correlation. [11] [5]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
30

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

“Best option” should be a constrained recommendation problem. Positive identification, rules, geography, time, command authority, civilian protection, system health, data quality, and reversibility should exclude impermissible choices before ranking. [11] [5]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
2
Open the full research route
31

EVIDENCE CAUTION

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

Uncertainty can multiply while the interface presents one deceptively precise score. Correlated sensor errors, stale tracks, domain shift, deception, and calibration drift can all survive into the final recommendation. [11] [5] [3]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
3
Open the full research route
32

EVIDENCE CAUTION

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

Human-in-the-loop and human-on-the-loop labels do not guarantee meaningful control when the operator lacks time, evidence, workload capacity, intervention reach, or freedom to reject the recommendation. [11] [5]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
2
Open the full research route
33

SUPPORTED BY SUBMITTED REPORTS

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

Autonomous swarms can act as edge nodes through local sensing, navigation, networking, and task allocation, but create risks of partitioned world models, compromised members, duplicate tasks, emergent behavior, and supervisory overload. [11] [5]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
34

EVIDENCE CAUTION

Use AI to preserve options and uncertainty, not to hide authority

Resilience, integrity, and autonomy

A person cannot meaningfully supervise a recommendation when the intervention path cannot reach the relevant function before the state becomes irreversible, even if the interface still shows an approval or abort control. [11] [5] [10]

How to read this state

This statement identifies a limit, failure mode, or interpretive hazard. It does not prove that every kill web contains the problem.

Retained reports
3
Open the full research route
35

SUPPORTED BY SUBMITTED REPORTS

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

The reports describe a structural mismatch: the legacy system is optimized for a bounded product with one sponsor, program office, baseline, contractor, budget line, test plan, and sustainment organization, while a kill web is a continuously changing system of systems. [9] [4]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
36

CROSS-REPORT ANALYSIS

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

The problem is not simply that acquisition law forbids agility. Existing pathways and authorities may be flexible, but they operate inside requirements, budgeting, test, cybersecurity, sustainment, and organizational systems that remain platform-centric. [9] [4]

How to read this state

This is a cross-report synthesis. It is not a direct quotation, a universal doctrinal definition, or proof that every implementation behaves this way.

Retained reports
2
Open the full research route
37

SUPPORTED BY SUBMITTED REPORTS

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

Interfaces, common data models, shared test environments, identity services, gateways, and integration software often lack the stable ownership and funding enjoyed by platform production and sustainment. [9] [4]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
38

SUPPORTED BY SUBMITTED REPORTS

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

Early data-rights and intellectual-property choices can determine whether the government can integrate, compete, upgrade, test, and sustain a system decades later. [9] [4]

How to read this state

One or more submitted reports support this statement. The label does not independently validate every citation inside those reports or establish a classified fielded configuration.

Retained reports
2
Open the full research route
39

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

Open architecture must be paired with conformance tests, interface governance, predictable transition funding, recurring competition, mission-outcome measures, secure software delivery, and incentives that reward replaceability rather than lock-in. [9] [4] [7]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
40

The unit of accountability is smaller than the unit of mission value

Acquisition, governance, and assurance

Replaceability is determined before and after contract award through mission ownership, delivered interface semantics, technical-data rights, executable conformance evidence, update authority, transition funding, and exit rights. [9] [4] [7]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
41

CROSS-REPORT ANALYSIS

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

The reports repeatedly conclude that a kill web is not one software product. It is a system-of-systems operating model that joins technology, data, people, policy, authorities, training, logistics, and test. [8] [1] [12] [9]

How to read this state

This is a cross-report synthesis. It is not a direct quotation, a universal doctrinal definition, or proof that every implementation behaves this way.

Retained reports
4
Open the full research route
42

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Assurance should test complete mission threads under simultaneous loss, latency, deception, stale data, identity failure, coalition boundaries, software rollback, and operator workload—not only nominal component performance. [1] [10] [11]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
43

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Authority must be explicit at each state: who may recommend, authorize, reserve, execute, interrupt, reassign, assess, and correct, and what evidence must accompany each transition. [11] [12] [13]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
44

PUBLICLY UNSPECIFIED

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Public descriptions often establish a program objective, demonstration, transition, or capability claim without disclosing fielded configuration, operating mode, error rate, authority setting, cyber design, or real-world performance. [1] [8] [11] [13]

How to read this state

The submitted public material does not establish this detail. Unknown must not be silently converted into either a reassuring or alarming claim.

Retained reports
4
Open the full research route
45

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Public explanation should preserve separate labels for official description, manufacturer claim, demonstration, independent corroboration, analyst interpretation, recommendation, unknown, and synthetic example. [2] [13] [1]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
46

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Mission-thread assurance should require state-to-evidence agreement across authority, identity, provenance, semantics, transport, human intervention, audit, recovery, and coalition release—not merely a successful end-to-end message exchange. [13] [10] [9]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
3
Open the full research route
47

A system of systems needs mission-thread assurance and visible unknowns

Acquisition, governance, and assurance

Mission-thread assurance should include semantic translation, role-tailored releasability, intervention effectiveness, and acquisition replaceability as explicit proof objects rather than background assumptions. [13] [12] [11] [9]

How to read this state

This is a recommended engineering, governance, procurement, or assurance control. It should not be read as evidence that the control is already deployed.

Retained reports
4
Open the full research route

INTERACTIVE / SYNTHETIC / FAIL-SAFE

Degraded-Mode Decision Drill

Choose a synthetic trust failure. The drill separates technical availability from evidentiary trust and decision authority. A function may remain online while its output is too stale, disputed, unauthenticated, or semantically uncertain to support the next consequential transition.

Synthetic condition

Nominal evidence and authority

Identity, time, provenance, semantics, transport, authority, and assessment are current and independently supported.

Permitted operating posture

The thread may remain composable under normal review. Availability still does not remove the need to check authority and evidence quality at each transition.

Next assurance action

Continue bounded monitoring and record the evidence state used for each transition.

Function Purpose Function status
01Observe Collect observations and preserve their origin, time, quality, and uncertainty. Available
02Normalize Translate formats without erasing units, classification, provenance, releasability, or ambiguity. Available
03Fuse Associate records while preserving disagreement and source dependence. Available
04Recommend Rank only eligible, compatible, sufficiently trusted options and expose abstention. Available
05Authorize Apply current identity, scope, policy, expiration, and command authority. Available
06Respond Carry out only the bounded, authorized, and technically supported function. Available
07Assess Independently verify the result, residual uncertainty, and need for recovery or rework. Available
Read every synthetic condition without JavaScript

Nominal evidence and authority

Identity, time, provenance, semantics, transport, authority, and assessment are current and independently supported.

Permitted operating posture: The thread may remain composable under normal review. Availability still does not remove the need to check authority and evidence quality at each transition.

Next assurance action: Continue bounded monitoring and record the evidence state used for each transition.

[13] [1] [12]

Cross-domain transport partition

Local observations remain available, but the path to remote fusion, recommendation, and independent assessment is intermittent.

Permitted operating posture: Do not claim a functioning cross-domain web. Preserve local awareness and reversible protective functions; hold consequential cross-domain composition until trustworthy transport or an approved degraded mode is restored.

Next assurance action: Prove route independence, message age, delivery completeness, and the authority of any local fallback.

[10] [3] [12]

Identity and authority service unavailable

Data and transport appear healthy, but current delegation, revocation, or operator identity cannot be authenticated.

Permitted operating posture: Observation and analysis may continue, but consequential authorization and response must fail closed. Connectivity is not permission.

Next assurance action: Restore a current authority chain or enter a preapproved safe state with explicit expiration and audit proof.

[10] [11] [13]

Time and sequence integrity disputed

Measurements are present, but clocks, message ordering, or evidence age disagree across nodes.

Permitted operating posture: Quarantine time-sensitive fusion and recommendation. A coherent display built from misordered data is not a trustworthy operating picture.

Next assurance action: Re-establish trusted time, preserve original timestamps, and replay the sequence against independent observations.

[10] [3] [7]

Semantic translation drift

Messages pass syntax checks, but units, confidence, object identity, releasability, or uncertainty may have changed at a gateway.

Permitted operating posture: Retain source-native records, expose the disagreement, and block cross-domain recommendation until round-trip semantics are demonstrated.

Next assurance action: Run conformance fixtures, compare original and transformed fields, and require independent corroboration.

[12] [8] [10]

Independent assessment path lost

The thread can observe, recommend, authorize, and respond, but no independent channel can verify the result or residual condition.

Permitted operating posture: Do not declare completion or automatically repeat a consequential action. Limit activity to what can be safely bounded without confirmation.

Next assurance action: Restore independent assessment, preserve unresolved state, and reconcile before reallocating scarce capacity.

[1] [13] [11]

Provenance conflict and correlated evidence

Several records appear to corroborate one another, but they derive from one underlying observation or a disputed transformation.

Permitted operating posture: Present alternative hypotheses and abstain from a confident recommendation. Route count is not source independence.

Next assurance action: Group dependent evidence, obtain an independent observation, and recalculate confidence without double counting.

[11] [10] [13]

SOURCE-PRESERVED RESEARCH

Retained report register

These are the thirteen submitted reports retained under the project documentation tree. Their presence supports traceability; it does not erase the qualifications or unknowns identified by the public synthesis.

  1. Submitted architecture report

    From Kill Chain to Kill Web: Operational Rationale, Technical Architecture, and the DARPA ACK Model

    Submitted architecture report. It distinguishes documented public facts, architectural synthesis, and recommended engineering targets; this release preserves those distinctions and does not independently reverify every cited source.

  2. Submitted concept and website report

    The Architecture of KillWebs.com: Operationalizing the Shift from Linear Chains to Dynamic Combat Networks

    Submitted website-strategy report. It informs public explanation, interaction design, and graph-oriented presentation; technical and historical claims remain subject to fresh verification before being stated as fact.

  3. Submitted defensive analysis

    The Vulnerability Surface of Kill Web Architectures: Non-Kinetic Disruption and Defense Strategies in Joint All-Domain Command and Control

    Submitted defensive-analysis report covering cyber, electromagnetic, timing, middleware, and zero-trust risks. The public synthesis retains only architecture-level defensive lessons and excludes exploit procedures.

  4. Submitted procurement analysis

    The Architecture of Friction: Why Legacy Defense Procurement Struggles to Build Kill Webs

    Submitted procurement analysis. It describes platform-centric incentives, budget fragmentation, open-architecture policy, and software acquisition using a mixture of sourced findings and interpretation.

  5. Submitted AI and autonomy analysis

    Algorithmic Warfare and the Evolution of the Kill Web: AI-Driven Decision-Making, Autonomous Swarms, and Strategic Stability

    Submitted AI-and-autonomy report. It organizes technical methods, legal questions, automation-bias risks, and strategic-stability concerns; numerical or program-specific claims require independent confirmation before reuse.

  6. Submitted joint-integration analysis

    The Algorithmic Battlespace: Architecting the Kill Web and the Future of Combined Joint All-Domain Command and Control

    Submitted JADC2 integration report. It supplies service, space, coalition, data-fabric, and mission-thread framing; scenario details are treated as illustrative rather than operational fact.

  7. Submitted architecture report

    The Architecture of Decision-Centric Warfare: Engineering the Transition from Kill Chains to Cross-Domain Kill Webs

    Submitted engineering architecture report. It covers modular hardware, edge computing, transport, middleware, capability orchestration, and centralization risk at a conceptual level.

  8. Submitted architecture report

    From Kill Chain to Kill Web: Strategic Rationale, Technical Architecture, and DARPA’s ACK Program

    Submitted executive architecture report. It is especially useful for the optionality argument, mission-specific timing, ACK boundaries, and unresolved public parameters.

  9. Submitted procurement analysis

    Why Legacy Defense Procurement Struggles to Build Kill Webs

    Submitted acquisition-reform report. It distinguishes statutory tools from the broader organizational, budgetary, data-rights, certification, sustainment, and incentive problems that shape a system of systems.

  10. Submitted defensive analysis

    Cybersecurity Vulnerabilities and Non-Kinetic Disruption of Kill Web Architectures

    Submitted defensive architecture report. It emphasizes identity, data integrity, time, supply-chain, electronic-warfare, AI, cloud, and insider risks, along with monitoring and resilience exercises.

  11. Submitted AI and autonomy analysis

    AI and Autonomy in Kill Webs

    Submitted AI decision-support report. It argues for uncertainty and provenance preservation, independent constraint gates, abstention, realistic human-machine testing, and bounded autonomy.

  12. Submitted joint-integration analysis

    Kill Webs as the Operational Engine of Joint All-Domain Command and Control

    Submitted joint-integration report. It treats JADC2 as the broader enterprise and the kill web as a composable mission thread spanning data, transport, authorities, services, partners, and effects.

  13. Submitted comparative analysis

    Kill Web vs. Kill Chain: From Sequential Targeting Processes to Composable Battle Networks

    Submitted comparative report. It provides the collection’s clearest process-versus-topology distinction and cautions that connectivity, resilience, command authority, and lawful use are separate variables.