Table of Contents
Search the Keywords & Search Queries column for the engineering difficulty, subject, or result you recognize. Each row explains the pattern’s contribution and links to its full body. The Readme offers selected starting examples; use the complete index for other working questions. The Dependencies column identifies results a method needs. Follow a dependency when its result is missing; you can start with a result already available.
Public units
| Unit | Reader use |
|---|---|
| Systems Engineering Principles Framework Readme | Follow connected methods from an engineering difficulty to a usable result and its changed conditions. |
| Agent work and support | Choose and use an available way to finish the task, select a worthwhile probe, or identify what to change when the way the work or its support is arranged keeps failing. |
| Citation | Cite this framework or one pattern with its author, title, release date, and publication address. |
| Systems Engineering Principles Framework Architectural Reference | Understand how 54 patterns connect common Systems Engineering, Platform Engineering, selected software Methods and engineering of agent work and support. |
| SYSE.Reference:1 | Choose a useful result and combine its needed contributions |
| SYSE.Reference:1.1 | The system-thinking long mantra in engineering |
| SYSE.Reference:2 | One System can appear in several engineering structures |
| SYSE.Reference:3 | Use and System concepts stay linked |
| SYSE.Reference:4 | Realization is recursive and continuous |
| SYSE.Reference:5 | Common platform work and software-specific Methods |
| SYSE.Reference:5.1 | Improve an agent’s choice and use of support |
| SYSE.Reference:5.2 | Construct and change a software delivery and recovery path |
| SYSE.Reference:6 | Configuration and evidence travel with continuing change |
| SYSE.Reference:7 | A Method has several structures; Work can be simultaneous |
| SYSE.Reference:8 | How project Work contributes to engineering culture |
| SYSE.Reference:9 | Architectural Rationale |
| SYSE.Reference:10 | Check the combined engineering answer |
| SYSE.Reference:11 | Scope and specialist boundaries |
| SYSE.Reference:12 | Intended use and non-use boundary |
| SYSE.Reference:13 | PatternID discipline |
| SYSE.Reference:14 | Source use and epistemic status |
| SYSE.Reference:15 | Shared engineering sources and architectural choices |
| SYSE.Reference:16 | Dependence on FPF and neighbouring products |
| SYSE.Reference:17 | Edition scope and refresh |
| SYSE.Sources:1 | Project-focus research sources |
| SYSE.Sources:2 | Professional sources and their use |
| SYSE.Application - Cross-Pattern Applications | Use two navigation walkthroughs and four worked applications in software, cyber-physical equipment and manufacturing. |
| Systems Engineering Principles Framework Sources and Edition Notes | Check professional source coverage, trace claims back to their sources, and find remaining profile obligations; edition scope and refresh explains the publication’s scope and reconsideration conditions. |
Part I - Project Focus, Environment, Consequences, and Problem/System-Family Development
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 1 | SYSE.1 - Choose and Reopen the Project System-of-Interest | Stable | Keywords: project system-of-interest, intended System, project focus, product, solution, project boundary, outside effect. Queries: “Which System is this project changing or bringing about?” “Why is this the useful focus now, and what could reopen it?” Establish a shared project-system choice that lets later use, architecture, realization, and assurance decisions refer to the same subject. | FPF A.1.SCR, C.11 |
| 2 | SYSE.16 - Recover the Systems and Conditions Needed for a Use Decision | Stable | Keywords: operational environment, suprasystem, using System, neighbours, interactions, operating conditions, context diagram. Queries: “Which surrounding Systems and conditions matter to this use decision?” “What containing, using, interacting, or boundary-crossing relations must the engineering account distinguish?” Recover the decision-relevant use context across its functional, constructive, interaction, and operating claims so the next concept or architecture choice has a grounded basis. | SYSE.1; FPF A.22 |
| 3 | SYSE.17 - Find Systems That May Bear Engineering Consequences | Stable | Keywords: affected Systems, consequences, bystander, future operator, maintainer, ecological System, direct relation, possible path. Queries: “Who or what may undergo a consequence of this engineering change?” “Which supported relation or possible path makes that bearer relevant?” Discover and qualify consequence-bearing Systems while the project can still change its concept, constraint, alternative, probe, or stop decision. | SYSE.1, SYSE.16; FPF A.1.CSD, C.28, A.10 |
| 4 | SYSE.2 - Develop Linked Use and System Concepts | Stable | Keywords: use concept, System concept, Concept of Operations, use case, scenario, user story, requirements, outside effect. Queries: “How will this System be used in a concrete situation?” “Which candidate System concept could support that use?” Link use claims to System-concept claims with their relevant neighbours, conditions, consequences, and any uncertainty; expose the next architecture question or the missing use premise. | SYSE.1, SYSE.16, SYSE.17; FPF C.17 |
| 5 | SYSE.54 - Make Domain Behaviour Explicit Together with EventStorming | Stable | Keywords: EventStorming, domain events, collaborative modelling, commands, policies, state, process modelling, domain behaviour. Queries: “Why do participants disagree about what happens and what should happen next?” “How can we expose missing decisions and test the behaviour before choosing software boundaries?” Use an existing account if it explains the behaviour the engineering work depends on. A rule or operating record may answer a remaining question. When people need to reconcile what happens, model events and test examples together. Identify any uncertainty that affects the intended use. | SYSE.2; FPF B.5.SD; A.3.3.TR when a state-change rule is needed |
| 6 | SYSE.22 - Coevolve Engineering Problems and System-Family Options | Stable | Keywords: problem formulation, System family, option generation, exploration, experiment, lineage, current alternatives. Queries: “How should the problem change when a prototype or operating observation reveals a new possibility?” “Which System-family options should remain under comparison now?” Develop problem formulations and possible System solutions together, preserving their lineage and returning new observations only to the choices they can change. | SYSE.1, SYSE.2, SYSE.6, SYSE.13; FPF C.11, C.17-C.19, G.11 |
Part II - Architecture, Descriptions, Offerings, and Professional Contributions
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 7 | SYSE.5 - Develop an Engineered System’s Functional Organization and Bearer Alternatives | Stable | Keywords: function, functional organization, bearer, allocation, interface, operating mode, placement, redundancy. Queries: “Which different organizations could produce the required outside effect?” “How could Systems share or alternate the functional contributions?” Develop functional and constructive alternatives together with their proposed allocations, interfaces, integration dependencies, and evidence, giving the architecture decision materially different structures to compare. | SYSE.2, SYSE.16; FPF A.6.F, A.22 |
| 8 | SYSE.6 - Decide and Reopen the Engineering Architecture | Stable | Keywords: architecture decision, functional allocation, module boundary, interface, control, redundancy, product family, trade-off, software architecture styles, layers, pipeline, plugins, microservices. Queries: “Which engineering architecture should the project choose for this use?” “Which evidence or changed condition would reopen that choice?” Understand software-style mechanisms where relevant; select the decision-relevant structures, compare materially different alternatives and accepted losses, and state the constraints that later realization and specialist decisions must respect. | SYSE.2, SYSE.5; FPF C.30, C.32 |
| 9 | SYSE.7 - Keep Engineering Descriptions Usable Together for Decisions | Stable | Keywords: engineering descriptions, model, requirement, CAD, simulation, digital twin, configuration, correspondence, contradiction. Queries: “Which claims across our models and records support this decision?” “Do they concern compatible subjects, configurations, intervals, and interpretations?” Maintain a usable ensemble of descriptions with explicit links to sources and returned evidence, justified correspondences, and identified contradictions that can change the engineering answer. | SYSE.6; FPF C.29, A.10 |
| 10 | SYSE.8 - Develop an Integrated Offering and Provider Concept for Using an Engineered Result | Stable | Keywords: integrated offering, provider concept, service, access, consumables, maintenance, support, financing, conditional payment, risk allocation. Queries: “What provider arrangement could fulfil this offering in its intended use?” “How do provider, access, and support choices change the technical concept and allocation of risk?” “How can we provide the result when payment depends on its technical acceptance?” Develop the offering and provider concept alongside the System concept, connecting the needed Systems, Work, promises, supported contributions and conditional finance. | SYSE.2; FPF A.2.3 |
| 11 | SYSE.9 - Use or Request Specialist Engineering Results | Stable | Keywords: specialist result, engineering contribution, receiving decision, safety, calculation, authority, evidence limits. Queries: “Which specialist answer could change this whole-System decision?” “What subject, configuration, conditions, and evidence must that answer address?” Use an adequate qualified specialist result or return its material limit. Select a new request only when its attainable contribution warrants the whole acquisition burden for the engineering choice; preserve the supplier Method and authority. | SYSE.2, SYSE.6; FPF A.15.9, C.11.DUA, A.6.REL, E.10.ROLE |
| 12 | SYSE.10 - Assess Research, Model, and Trial Results for an Engineering Decision | Stable | Keywords: research result, engineering model, simulation, experiment, prototype, field trial, formal proof, validity. Queries: “What can this research, model, or trial result support in our engineering decision?” “Which implementation correspondence or untested condition limits that use?” Assess what the result supports, its remaining uncertainty, and any need for further observation, so the decision uses partial knowledge within its justified limits. | SYSE.6, SYSE.7; FPF A.10, C.16 |
Part III - Obtaining Engineering Results, Recursive Realization, Platforms, Independent Constituents, and Evolvability
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 13 | SYSE.24 - Choose How the Project Will Obtain a Needed Engineering Result | Stable | Keywords: make or buy, reuse, supplier, outsourcing, provider, AI assistance, integration burden, migration, exit, platform obtaining, repair, retain, share, build or buy. Queries: “How can the project obtain the same needed engineering result in different ways?” “Which arrangement includes the assurance, support, capability, and exit costs?” Use an adequate existing engineering or acquisition comparison, or compare the complete arrangements still in question. The result can be a choice, retained alternatives, rejection, a worthwhile probe or the question that prevents comparison. | SYSE.1, SYSE.2; FPF C.11, C.18 |
| 14 | SYSE.3 - Plan the Next System-Realization Action (Recursive Realization Network) | Stable | Keywords: realization network, builder System, fabrication, adaptation, installation, qualification, build the builder, feasibility. Queries: “Which Systems and Work can realize the selected architecture?” “Which realization branch, if any, lacks support, including the means to build or change a builder?” Use a sufficient existing realization plan. If a branch lacks support, identify the next feasible action, when to stop that inquiry and which earlier decision to revisit if the branch proves infeasible. | SYSE.6, SYSE.8; FPF E.18.NET, A.15 |
| 15 | SYSE.11 - Integrate a System for One Bounded Use | Stable | Keywords: integration, bounded use, usable increment, installed interface, configuration, operating evidence, fallback. Queries: “Which configuration is usable for the intended purpose now?” “What evidence shows whether the parts or changes work together?” Use a qualified assessment covering the configuration and intended use. If more work is needed, integrate any changes still required and assess the resulting System from suitable observations, obtaining further evidence where necessary. Return the result with its grounds, limits and fallback. | SYSE.3, SYSE.12, SYSE.13, SYSE.18; FPF A.3.4 |
| 16 | SYSE.12 - Develop or Retain an Engineering Platform for Practitioner Work | Stable | Keywords: engineering platform, practitioner Work, model repository, manufacturing cell, test bench, AI assistant, evidence store, support, platform development, readiness, later use. Queries: “What must enable the engineers’ modeling, building, testing, release, and observation Work?” “Which platform arrangement helps one Work family and obstructs another?” Use an adequate readiness or use assessment to decide whether to keep the arrangement or change it. For a readiness decision, use the evidence available before the work. For a decision about actual use, use observations of the platform’s contribution to practitioner tasks and the engineering result. | SYSE.15, SYSE.20; FPF A.2.2 |
| 17 | SYSE.18 - Decide Whether and How to Integrate Systems Governed by Different Agents | Stable | Keywords: independent constituents, system of systems, interoperability, authority, interface, commitment, independent evolution, platform federation. Queries: “How can Systems governed by different Agents work together for this use?” “Which contribution or interface may change outside the integrating project’s control?” Identify who governs the constituent changes that could affect this use. Use the existing coordination decision when its authority, conditions and evidence still cover the use. If a choice remains, compare the alternatives and identify any agreements or observations each requires. | SYSE.6; FPF E.18.NET |
| 18 | SYSE.23 - Choose What to Change So Later System Changes Become Easier | Stable | Keywords: evolvability, future change, modularity, configurable family, builder platform, change lead time, test capability. Queries: “Can we retain the current arrangement for the changes we expect, or would changing it make them easier?” “Is the limiting condition in the project System, its builders, the platform, or their joint arrangement?” Compare retaining the current arrangement with changes to the product, its builders or both. Use the future changes, dependencies, burden and evidence for their expected contribution to decide whether an investment is warranted. | SYSE.22, SYSE.3, SYSE.6, SYSE.12, SYSE.20; FPF C.11, C.25, C.30, C.32 |
Part IV - Configuration, Continuing Change, and Source Continuity
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 19 | SYSE.13 - Establish Configuration Identity, Variants, and Effectivity | Stable | Keywords: configuration identity, unit, variant, version, baseline, effectivity, installed state, bill of material. Queries: “Which actual or proposed configuration does this claim concern?” “For which units, conditions, and intervals does the evidence apply?” Use an existing configuration basis when it identifies the subject, conditions and interval with enough evidence for the decision. Establish or reconcile any missing identity, configuration or applicability across the descriptions the decision needs. | SYSE.6; FPF A.22, C.2.1, C.27.TA |
| 20 | SYSE.14 - Make a Release Decision for Named Engineering Work or Use | Stable | Keywords: engineering change, release, release for build, release to service, configuration, permission, evidence continuity. Queries: “What may be released for which named Work or use?” “Which evidence and authority support that release for the named use?” Use an existing release result when it covers the release subject, next use, supporting evidence, deciding authority and any required permissions. Otherwise connect the missing grounds and return the bounded decision or the premise that blocks it. | SYSE.13, SYSE.19; FPF C.11, A.10, B.3 |
| 21 | SYSE.19 - Revalidate Engineering Decisions When a Relied-on Source Changes | Stable | Keywords: source change, supplier revision, standard, research correction, relied-on premise, impact analysis, revalidation. Queries: “Which engineering decisions relied on the claim that changed?” “Does the change affect current use, configuration, or release evidence?” Compare the source claims. A compatible republication can finish with any needed reference repair. Trace material changes to the engineering uses that depend on them, preserve earlier editions and decisions, and revalidate the affected answers when needed. | SYSE.7, SYSE.10; FPF C.2.1, A.10, G.11 |
Part V - Assurance, Method and Work Architecture, Repertoire, and Cultural Continuation
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 22 | SYSE.4 - Decide Whether and How to Challenge an Engineering Claim | Stable | Keywords: engineering assurance, claim, challenge, test, measurement, inspection, evidence use, validity boundary. Queries: “Which challenge could change reliance on this engineering claim?” “What configuration, conditions, uncertainty, and time limit the returned evidence?” Establish what the available evidence supports. Use a sufficient result, narrow or stop unsupported reliance, or select a challenge when its attainable contribution warrants the work. If reliance changes, reassess the affected focus, concept, architecture, realization or specialist answer. | SYSE.10, SYSE.11, SYSE.14; FPF A.10, B.3 |
| 23 | SYSE.15 - Choose and Refresh the Engineering Methods Needed by a Project | Stable | Keywords: engineering Method, repertoire, situation, application profile, variant, trial, qualification, refresh. Queries: “Which engineering Methods are needed to obtain this project’s results?” “What evidence would justify keeping, changing, or rejecting a Method variant?” Select and refresh the repertoire for the System and engineering situation, retaining specialist Methods and the distinction between engineering and the System’s operating Methods. | FPF A.3.2, C.36, E.23 |
| 24 | SYSE.20 - Reconcile Overlapping Engineering Work and Required Order | Stable | Keywords: overlapping engineering Work, concurrency, required order, result dependency, Method structure, configuration, coordination. Queries: “Which modeling, implementation, trial, and review Work can overlap?” “Which result or released configuration must exist before another use?” Reconcile the relevant Method, Work, System, representation, and support structures, to establish required order or permitted overlap, or assess whether an arrangement meets the specified use. When a choice is requested, compare the options and return the authorized outcome or the remaining gap. | SYSE.15; FPF A.15.1, A.22.CGUS, E.18.NET |
| 25 | SYSE.21 - Decide Whether and How to Change Systems Engineering Culture | Stable | Keywords: Systems Engineering culture, Method variant, practitioner population, transmission, retention, branching, intervention. Queries: “Should this engineering practice continue, change, branch, or stop across the relevant population?” “What evidence distinguishes visibility, local enactment, and wider continuation?” Recover which method practitioners enacted, any supported variant and the cultural relations for a bounded population and period; return the qualified current account or continuation choice, selecting further inquiry or intervention only when warranted. | SYSE.20; FPF C.36, G.11 |
Part VI - Platform Engineering: Supported Use and Continuing Change
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 26 | SYSE.25 - Decide Whether and How to Improve a Platform from Practitioner-Task Evidence | Stable | Keywords: platform improvement, practitioner tasks, failed attempts, transferred burden, competing hypotheses. Queries: “Which platform difficulty should we address next?” “What observation could reject this proposed improvement?” Use task evidence for one user population, including failed or assisted attempts that can change the choice. Compare improvement mechanisms and the burden for users and providers. Return a bounded choice to retain or change provision, a discriminating probe or the specific missing premise; use SYSE.24 when complete obtaining arrangements compete. | SYSE.12; SYSE.24 for whole obtaining; SYSE.36 for software task measures |
| 27 | SYSE.26 - Design a Supported Platform-Use Path | Stable | Keywords: supported use, request and result, variants, retry, cancellation, support. Queries: “What must the user supply to obtain a usable engineering result?” “What is safe after an interrupted request?” Use an adequate existing interaction or design one that connects required inputs, waiting, completion, failure and capable support. Return the supported-use description and any trial evidence, worthwhile trial or missing provision the decision needs; distinguish description, actual operation and readiness to rely on it. | SYSE.12, SYSE.13; SYSE.8 if the provider concept is missing |
| 28 | SYSE.27 - Evolve Platform Interfaces and Contribution Paths | Stable | Keywords: interface evolution, compatibility, units, defaults, physical connections, adapters, contribution, versioned change. Queries: “Does this change preserve the consumer’s promised result?” “How can a contributor test and maintain an extension?” Compare implementation change, extension, adapter, explicit break and refusal by their effects on the behavior consumers rely on. Return a bounded evolution decision with consumer effects, a way to test the contribution and the maintenance/support arrangement or gap; send required migration to SYSE.29. | SYSE.13, SYSE.26; OCE.4/OCE.6 when organizational results are missing |
| 29 | SYSE.28 - Place Qualified Controls in a Supported Platform-Use Path | Stable | Keywords: control placement, evidence reach, subject identity, reuse, unavailable checker, authorized exception. Queries: “Where can this property change before it is relied on?” “What stops when the required check is unavailable?” For a supplied constraint, trace where the relevant property can change, then place verification or enforcement to protect the dependent action. Return the placement, outcome meanings, applicability and remaining gaps, preserving separate domain qualification and decision authority. | SYSE.4, SYSE.13, SYSE.26; SYSE.32 for artifacts; applicable control authority |
| 30 | SYSE.29 - Decide Whether and How to Migrate or Retire a Supported Platform Path | Stable | Keywords: platform migration, retirement, retained state, infrequent consumers, coexistence, recovery. Queries: “Can users complete the required work through the receiving path?” “What must remain after the old interface stops serving?” Recover affected uses and obligations and compare transition choices. Qualify a selected increment and its failure response, reusing adequate evidence. Return a bounded decision, its next increment when needed, or the missing premise, with remaining uses and obligations accounted for. | SYSE.13, SYSE.18, SYSE.27; SYSE.34 for software/data |
| 31 | SYSE.53 - Resolve a Usable Composition of Independently Released Components | Stable | Keywords: independent releases, package composition, typed dependencies, shared meaning, isolation, resolver, conflict, unknown premise. Queries: “Which available releases can support this joint use?” “What must change when the preferred update conflicts?” Derive requirements and sharing scopes from what the joint use consumes, obtain a supported selection or an explained conflict or unresolved premise, and return exact inputs to realization. | SYSE.13, SYSE.27; SIE.2 / SIE.4 / SIE.6 for semantic premises; CMP.4 for search; SYSE.30 / SYSE.33 for realization |
Part VII - Software Platform Engineering
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 32 | SYSE.30 - Make Software Builds Repeatable and Traceable | Stable | Keywords: software builds, resolved inputs, clean execution, byte reproducibility, timestamps, caches. Queries: “Which inputs produced the tested artifact?” “Can the specified package be reconstructed byte for byte?” Recover the build, its inputs and the specified artifact. For a build procedure, use qualified evidence that it produces the output under the stated conditions. For reproducibility, compare independent executions and obtain any missing evidence. Return the procedure, artifact-to-build account or qualified comparison needed for that question, with any missing input or unexplained difference. | SYSE.13, SYSE.31, SYSE.33 for qualified inputs and conditions |
| 33 | SYSE.31 - Keep Software Feedback and Continuous Integration Fast and Trustworthy | Stable | Keywords: software feedback, test meaning, flaky failures, quarantine, continuous integration, mainline. Queries: “Does this green check answer the current engineering question?” “Why do passing branches fail when combined?” Bind checks to the exercised artifact and conditions, preserve necessary slow or human evidence, and direct failures to those who can correct them. Return a usable feedback arrangement, or use the named DORA CI Method to obtain a checked shared-mainline revision for an already testable change awaiting integration. | SYSE.30, SYSE.33; DORA CI Method when integration is missing |
| 34 | SYSE.32 - Promote Verified Artifacts without Rebuilding | Stable | Keywords: artifact promotion, artifact identity, provenance, consumer trust, destination configuration, verification. Queries: “Does the received package still carry the tested artifact’s evidence?” “What does a valid signature leave unestablished?” Preserve the selected artifact across transfer, separate destination configuration and apply the named SLSA verification Method with the consumer’s configured expectations when authenticity is required. Return an applicable promotion basis or a specific identity, evidence or trust gap; runtime deployment remains a separate result. | SYSE.13, SYSE.28, SYSE.30, SYSE.31; SLSA v1.2 verification |
| 35 | SYSE.33 - Provide Reconstructible Software Development and Test Environments | Stable | Keywords: development environment, test environment, reconstruction, infrastructure as code, isolation, state reconciliation, cleanup. Queries: “Can another user reconstruct an environment that supports this task?” “What should happen after partial provisioning or a failed retry?” Specify the resources, data, access and operating conditions the task needs. Use an environment already qualified for that task, or reconstruct and exercise one to establish its suitability. The first result is an observed usable environment with its limits, or a reconstructible description with the precise missing resource, permission or data source. | SYSE.13, SYSE.26; SYSE.34 for persistent state |
| 36 | SYSE.34 - Change Software and Persistent Data with a Recovery Boundary | Stable | Keywords: persistent data, schema migration, compatible coexistence, concurrent writes, backfill, idempotence, recovery, information loss. Queries: “Can old and new software safely use the same changing data?” “What remains recoverable after a partial or lossy migration?” Establish whether the change affects data meaning or compatibility. An unchanged contract may need only a compatibility check. Where a transition is needed, identify the writers and intermediate states and qualify the transition and its promised recovery actions against the histories and partial effects they must support. Return the applicable compatibility result or the bounded procedure with its operating limits, decision holder and any unresolved conditions. | SYSE.13, SYSE.27, SYSE.33; engine-qualified change/recovery Methods |
| 37 | SYSE.35 - Control Software Release Exposure with User-Relevant Signals | Stable | Keywords: release exposure, canary, representative tasks, control population, attribution, delayed outcomes, feature activation, recovery. Queries: “What evidence justifies widening use of a deployed candidate?” “When is a canary inconclusive rather than successful?” Select the population, task, observation interval and authorized exposure; compare outcomes while accounting for shared effects and state compatibility. The first result defines the comparison, decision conditions and qualified recovery action; observations then support a bounded proceed, stop or inconclusive result. | SYSE.34, SYSE.36, SYSE.41; SYSE.14 release authority |
| 38 | SYSE.36 - Measure User Tasks and Set Software-Service Reliability Objectives | Stable | Keywords: user task, service level indicator, service level objective, SLI, SLO, eligibility, missing outcomes, error budget. Queries: “Does this measurement count successful user tasks or only healthy components?” “What denominator and objective should govern this service?” Define eligible attempts, what counts as a good result for the task, and observation gaps before choosing telemetry; when an SLO is needed, agree the objective and response policy with the relevant users, providers and decision holders. The first result is a task measurement definition and exercised instrumentation, or a measurement gap. | SYSE.4; subject owner, task meaning and suitable instrumentation |
| 39 | SYSE.37 - Alert on Actionable Software-Service Reliability Risk | Stable | Keywords: actionable alert, error budget, burn rate, multiple windows, sparse traffic, missing data, response capacity. Queries: “Does this reliability signal require action soon enough to justify an alert?” “Will the alert reach someone able to respond?” Choose an alert rule for actionable task failure, budget risk or lost observation, deriving budget-risk rules from the matching task, objective and consumption. Test evaluation, delivery, acknowledgement and the permitted response together. The first result is an exercised alert rule with its response and blind spots, or a precise observation or response gap. | SYSE.36; SYSE.38 for response; responder’s authority for the proposed action |
| 40 | SYSE.38 - Diagnose and Restore a Failed Software-Platform Task | Stable | Keywords: failed platform task, diagnosis, incident response, safe mitigation, partial effects, competing causes, restoration, specialist request. Queries: “Which stage of this user attempt failed, and what can restore useful work now?” “What observation would distinguish the remaining causes?” Establish the impact and what is known about the attempt. Recover an existing result or use a qualified repair when it suffices. Where needed, distinguish plausible causes and choose a permitted mitigation that preserves the required result and dependent effects in every still-possible state. The first result is a restored task or usable limited result, or a specific question to a specialist about an unresolved cause or effect. | SYSE.26, SYSE.34, SYSE.36, SYSE.40, SYSE.41; OPS.3/OPS.4 when needed |
| 41 | SYSE.39 - Decide Whether and How to Reduce the Total Burden of Repetitive Platform Work | Stable | Keywords: repetitive platform work, toil, automation, task simplification, transferred burden, maintenance, exceptions, horizon. Queries: “Will this change reduce total work or move it to users?” “When does automation repay its construction and operating cost?” Compare retaining the method, simplifying the task, removing its cause, improving the supported interaction and automation using user effort, provider effort, waiting and risk over a stated horizon. Return a justified choice or identify what prevents a choice on the available evidence. For a selected intervention, state how total burden will be compared after use. | SYSE.25, SYSE.26, SYSE.36; SYSE.24 for whole obtaining |
| 42 | SYSE.40 - Protect Software Platform Capacity and Isolate Failure | Stable | Keywords: platform capacity, workload classes, admission control, bounded queues, isolation, deadlines, retry budget, serialization, recovery capacity. Queries: “Which shared limit lets one workload delay or disable another?” “Can the platform recover while demand remains excessive?” Use an existing qualified arrangement where it covers the demand and failure conditions. Otherwise compare capacity and execution changes and qualify the relevant admission, isolation, waiting or retry mechanisms across shared dependencies. The first result is a tested protection arrangement for a stated envelope, with observed admitted, delayed and refused work and recovery limits, or the conditions that remain unqualified. | SYSE.26, SYSE.33, SYSE.36; SYSE.34/SYSE.41 for stateful recovery |
| 43 | SYSE.41 - Deploy a Verified Software Artifact with the Target Configuration and Handle Partial Failure | Stable | Keywords: verified artifact, deployment, effective configuration, runtime readback, dependency test, partial failure, unknown effects, recovery. Queries: “Did the intended software and configuration become usable on each target?” “What may be retried after an interrupted deployment?” Use applicable observations and tests of the existing runtime. Where a change is needed, carry out the authorised deployment steps that are safe from the possible partial states. The first result is an observed runtime configuration with its test, interval and limits, or a precise failed or unknown partial state; obtain the required authority before any action that exposes live work or grants a release. | SYSE.13, SYSE.32, SYSE.33, SYSE.34; SYSE.11/SYSE.35 as receivers |
Part VIII - Engineering Agent Work and Support
| § | ID & Title | Status | Keywords & Search Queries | Dependencies |
|---|---|---|---|---|
| 44 | SYSE.42 - Use a Selected Tool and Apply Its Result | Stable | Keywords: selected aid, input binding, calculator, invocation, uncertain effect, receiving use. Queries: “Did we apply the tool to the intended input?” “How does its result enter the task?” Use a sufficient result for the intended input when one is available. Otherwise bind the input, perform the supported operation and apply its qualified result. Use the interface contract to resolve an uncertain effect or establish which continuation is safe. | C.38/C.11; A.15.7 / A.15.9; C.24; SYSE.26–SYSE.28 / SYSE.43 / SYSE.46 / SYSE.47 / SYSE.52 |
| 45 | SYSE.43 - Maintain External Memory for Continuing Work | Stable | Keywords: worksheet, episode, provenance, conditional index, retrieval, reorganization, retirement. Queries: “Which relation did the record lose?” “Should these episodes share a summary?” Use the available record or context when it supports the continuation. Where it falls short, locate the loss and repair recording, retrieval or use. Test a proposed reorganization against the source episodes and available relevant contrary experience. | A.15.8; A.15.8.MS; SYSE.42 / SYSE.46 / SYSE.47 / SYSE.48 / SYSE.52 |
| 46 | SYSE.44 - Divide and Recombine Work across Agents | Stable | Keywords: participant brief, contribution, shared premise, dependency, integration, source error. Queries: “Does this division repay its handovers?” “Which return became stale?” Check whether one performer can supply the whole result under the required conditions and authority. Compare the feasible arrangements. When separate contributions are required or worthwhile, obtain or reuse them, reconcile their results and reconsider the conclusions affected by changed premises. | A.15.9; C.38/C.11; SYSE.9 / SYSE.18 / SYSE.20 / SYSE.46 / SYSE.47 |
| 47 | SYSE.45 - Train an LLM Policy from Qualified Interaction Experience | Stable | Keywords: policy training, adapters, demonstrations, feedback, distillation, retained support. Queries: “Which stable contribution should the policy learn?” “What must remain externally available?” Use an applicable comparison to decide whether to keep the policy and its support or try a training change. For a trial, qualify the experience, select the learning mechanism and compare the resulting configuration with the unchanged baseline. | CMP.7; E.23.CDI; SYSE.49 / SYSE.46; SYSE.47 for a fixed-model alternative |
| 48 | SYSE.46 - Test an Agent and Its Support in Representative Work | Stable | Keywords: performing arrangement, matched support regimes, selection, result use, restraint, burden, persistence. Queries: “What can this configuration achieve on the intended task?” “Does support help, and does the agent choose it well?” Use or obtain bounded evidence with input, action, return and use distinguished; test delay or shift for the reliance claimed. | SYSE.10 / SYSE.4 / SYSE.33 / SYSE.49; A.15.8; E.23.CAE; HCD.12 / HCD.13 |
| 49 | SYSE.47 - Construct and Revise an Agent’s Execution Procedure | Stable | Keywords: operative sequence, consumed return, transition, written procedure, controller, recovery, stop. Queries: “Why does available material fail to advance the task?” “What sequence can this performer enact?” Keep an adequate procedure, or connect the supplied operations through usable inputs, progress, recovery and completion. | SYSE.25 / SYSE.26–SYSE.28 / SYSE.40 / SYSE.42–SYSE.46 / SYSE.48–SYSE.52; C.32.MWA.PP and B.1.5.CM; MMP.8.SD / MMP.17 |
| 50 | SYSE.48 - Construct and Maintain Reusable Tools and Procedures | Stable | Keywords: candidate operation, worksheet, parameterization, new input, defeating case, discovery, variant. Queries: “Which operation can be recovered from these attempts?” “Will the aid work on another input?” Compare the aid with direct work. Keep an adequate approach or construct and qualify the needed aid, make it discoverable and consolidate only demonstrated overlap. | A.15.8.MS and A.3.1.MV; F.0.2.MA for allocation; SYSE.26 / SYSE.27 / SYSE.39 / SYSE.42 / SYSE.46 / SYSE.47 / SYSE.49 / SYSE.52 |
| 51 | SYSE.49 - Construct Informative Tasks and Feedback Environments for Agent Work | Stable | Keywords: informative experience, practice material, fixture, feedback, shared error, transfer. Queries: “Which interaction exposes this failure?” “Could the task and its judge share a mistake?” Use a suitable existing environment or construct the missing tasks, working conditions and feedback for the receiving question. Qualify the experience for that use. | HCD.6 / HCD.7 / HCD.11–HCD.13; SYSE.33 / SYSE.45–SYSE.48 / SYSE.50 / SYSE.51; MMP.17; SYSE.10; CMP.7 |
| 52 | SYSE.50 - Construct and Calibrate an Agent’s Assistance Policy | Stable | Keywords: assistance, accessible signal, benefit, need, confidence, calibration, fallback. Queries: “Which observable condition should change the help decision?” “Can an aid be worthwhile despite unaided ability?” Use the evidence to retain a suitable assistance rule or construct and test one for this performer, task and cost of help. | C.38/C.11; A.15.7 / A.15.9; SYSE.45–SYSE.47 / SYSE.49 / SYSE.51 / SYSE.52; HCD |
| 53 | SYSE.51 - Construct Adaptive Control of Reasoning and Tool Effort | Stable | Keywords: further reasoning, tool effort, time, samples, branches, completion reserve, early stop. Queries: “Which further move can improve this result?” “Can we still finish after it?” Use a sufficient fixed or manual allocation, or construct an adaptive rule and compare whole-task results and total effort with feasible alternatives. | C.11 / C.11.DUA; CMP.4; SYSE.39 / SYSE.40 / SYSE.45–SYSE.47 / SYSE.49 / SYSE.50 / SYSE.52 |
| 54 | SYSE.52 - Prepare Working Material and Tools for an Agent’s Next Step | Stable | Keywords: readable layout, carry, working input, summary, pointer, evidence loss, tool view, recovery. Queries: “What must the next step receive?” “Did selection hide a decisive condition?” Use material and tools that already support the next step, or prepare what is missing. Check uncertain recovery and restore a consequential omission or report its limit. | SYSE.42 / SYSE.43 / SYSE.46 / SYSE.47 / SYSE.50 / SYSE.51; EXD.1 / EXD.3 / EXD.6; HCD |