Skip to content
CHEREngineering workspace
Experience
Appearance

Generated model

Programme status

Rendered as produced. Every figure below was measured when the model was generated; this screen computes none of them and completes none of them.

Read-only. Everything here is read from what CHER has recorded. Nothing on this screen decides, approves, records an outcome or writes to governance, evidence or audit.

Status model

source GENERATED_STATUS_MODEL

Every rung, ratio and segment in this model was produced by running the checks at generation time. A screen renders it as given. It may not derive a completion figure, a percentage or a rung of its own.

Schema
cher.programme-status/1
Commit
775ea10
Generated at
2026-09-05 20:38 UTC
Plane B measured
TRUE
Packages conforming
20 of 21
Validation rules executable
6 of 18
Check counts
FAIL 0 · KNOWN 4 · NOT RUN 0 · PASS 24 · STALE 0
Panels
14

Validation rules and their executables

A rule identifier and the check that runs under it are separate facts and are shown as separate fields. An identifier with an executable beside it does not by itself say the executable checks what the identifier governs.

  • VAL-AUDIT-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-AUTHZ-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-CHER-AGENT-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-CHER-AUTHORITY-001

    EXECUTED
    • executed by 09_Factory_Bootstrap/plane-b/verify_decision_authority.py · does not cover: auditing of refused attempts, which a database trigger cannot do because a refused write rolls back its own audit row, and which belongs to the unbuilt API layer the rule is scoped to; and the application identities in cher.actor, since no application layer exists to hold them

  • VAL-CHER-GEO-001

    EXECUTED
    • executed by 04_Canonical_Contracts/tooling/cher_validate.py · does not cover: whether the engine, evidence store and audit log resolve to infrastructure in the home jurisdiction, which requires a deployment to inspect and nothing here inspects

  • VAL-CHER-PLANE-001

    EXECUTED
    • executed by 09_Factory_Bootstrap/plane-b/verify_console_boundary.py

  • VAL-CHER-SNAPSHOT-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-CHER-TENANT-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-CHER-TRACE-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-CHER-TRUTH-001

    EXECUTED
    • executed by 04_Canonical_Contracts/tooling/cher_validate.py · does not cover: factory results at rest, which check_execution_result.py covers at acceptance, and whether an evidence substrate can support the integrity class it is given, which nothing inspects
    • executed by 09_Factory_Bootstrap/compiler/check_execution_result.py · does not cover: package-side provenance and evidence contracts, which cher_validate.py covers, and whether a substrate supports its claimed integrity class, which nothing inspects
    • executed by 09_Factory_Bootstrap/plane-b/verify_evidence_ingestion.py · does not cover: the package-side provenance and evidence contracts, which cher_validate.py covers. This covers the ingestion path: what may and may not be written when a check result becomes a record
    • executed by 09_Factory_Bootstrap/plane-b/verify_evidence_integrity.py · does not cover: package-side provenance and evidence contracts, which cher_validate.py covers, and factory results at acceptance, which check_execution_result.py covers. This covers the stored path only: what the evidence store will and will not persist

  • VAL-CHER-WORKER-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-PERF-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-PURPOSE-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-RETENTION-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-SECRETS-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-TENANT-001

    NOT RUN
    • executed by: nothing recorded

  • VAL-TRACE-001

    EXECUTED
    • executed by 04_Canonical_Contracts/tooling/cher_validate.py · does not cover: whether factory output conforms to the package, which check_execution_result.py covers for a submitted result
    • executed by 09_Factory_Bootstrap/compiler/check_execution_result.py · does not cover: requirement-level upstream and downstream links, which cher_validate.py covers on the package

  • VAL-XBORDER-001

    EXECUTED
    • executed by 04_Canonical_Contracts/tooling/cher_validate.py · does not cover: whether a recorded ground is supported by evidence, which this module does not fetch; the rule is declared MANUAL_REVIEW and the automated part of it is the declaration check only

00 · We are here: the north-star lifecycle

State is computed from PRODUCES declarations, from whether verify_pack.py runs the declaring module, and from the ratios in panel 03. The spine is the target architecture.

  • checkStates
    {}
    id
    BUSINESS_INTENT
    label
    Business intent
    note
    A person's words. The input to the pipeline, and not a component.
    qualifier
    NOT_APPLICABLE
    reachesLiveSystem
    false
    refs
    {"check":null,"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":null,"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/01_CHER_Concept_and_Context.md","task":null}
    rung
    SPECIFIED
    specifiedIn
    02_Target_Architecture/01_CHER_Concept_and_Context.md
  • checkStates
    {"context inference":"PASS"}
    checks
    context inference
    exercisedByTheRunner
    09_Factory_Bootstrap/plane-b/verify_context_inference.py
    id
    CONTEXT_INFERENCE
    label
    Context inference
    note
    Intent to a solution context profile. TSK-CHER-003.
    producedBy
    {"doesNotCover":"the conversation. This takes one statement of intent plus optional answers and produces one profile; there is no dialogue loop, no re-asking, and no model. It also does not choose knowledge, which is Knowledge Resolution's answer against a named snapshot","module":"04_Canonical_Contracts/tooling/infer_context.py"}, {"doesNotCover":"the conversation and any use of a model. This proves the deterministic core: ordinary language to a governed profile, and that profile driving snapshot-bound resolution","module":"09_Factory_Bootstrap/plane-b/verify_context_inference.py"}
    reachesLiveSystem
    true
    refs
    {"check":["context inference"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["04_Canonical_Contracts/tooling/infer_context.py","09_Factory_Bootstrap/plane-b/verify_context_inference.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/03_Context_Inference_Layer.md","task":["TSK-CHER-003"]}
    rung
    EVIDENCED
    specifiedIn
    02_Target_Architecture/03_Context_Inference_Layer.md
  • checkStates
    {"applicability resolution":"KNOWN","corpus diff":"PASS","impact assessment":"PASS","snapshot resolution":"PASS"}
    checks
    applicability resolution, corpus diff, impact assessment, snapshot resolution
    exercisedByTheRunner
    08_Knowledge_Repository/tooling/resolve_applicability.py, 09_Factory_Bootstrap/plane-b/verify_corpus_diff.py, 09_Factory_Bootstrap/plane-b/verify_impact_assessment.py, 09_Factory_Bootstrap/plane-b/verify_snapshot_resolution.py
    id
    KNOWLEDGE_RESOLUTION
    label
    Knowledge resolution / compiler
    note
    Which knowledge applies, what changed, and what a change means for what was already issued. TSK-CHER-002.
    producedBy
    {"doesNotCover":"the decision. It computes which issued packages a knowledge change reaches, classifies the change and reports what their declared basis cannot support; whether an affected package is re-issued, superseded or left alone is a governed decision under INV-CHER-007 and nothing here reaches it","module":"08_Knowledge_Repository/tooling/assess_impact.py"}, {"doesNotCover":"behaviour three, the impact of a knowledge change on packages already issued. This establishes that knowledge changed and whether a supplied context resolves differently; it reaches no conclusion about an issued package, which is a governed decision","module":"08_Knowledge_Repository/tooling/diff_snapshots.py"}, {"doesNotCover":"the corpus diff and the impact of a knowledge change on issued packages, which are behaviours two and three of the scope document. Behaviour one is complete: resolution reads an immutable snapshot, is attributable to it and reproduces from it","module":"08_Knowledge_Repository/tooling/resolve_applicability.py"}, {"doesNotCover":"behaviour three. This proves behaviour two: that a corpus difference is computed deterministically, attributed, and recorded. It reaches no conclusion about any issued package","module":"09_Factory_Bootstrap/plane-b/verify_corpus_diff.py"}, {"doesNotCover":"the governed decision itself. This proves behaviour three computes the impact set, classifies the change, finds what a declared basis cannot support, and stops there","module":"09_Factory_Bootstrap/plane-b/verify_impact_assessment.py"}, {"doesNotCover":"the corpus diff and the impact of a knowledge change on issued packages, which are behaviours two and three of the scope document. This covers behaviour one only, resolved from an immutable snapshot","module":"09_Factory_Bootstrap/plane-b/verify_snapshot_resolution.py"}
    ratio
    3 of 3 behaviours the scope declares
    ratioTruthClass
    MEASURED
    reachesLiveSystem
    true
    refs
    {"check":["applicability resolution","corpus diff","impact assessment","snapshot resolution"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["08_Knowledge_Repository/tooling/assess_impact.py","08_Knowledge_Repository/tooling/diff_snapshots.py","08_Knowledge_Repository/tooling/resolve_applicability.py","09_Factory_Bootstrap/plane-b/verify_corpus_diff.py","09_Factory_Bootstrap/plane-b/verify_impact_assessment.py","09_Factory_Bootstrap/plane-b/verify_snapshot_resolution.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"08_Knowledge_Repository/02_Knowledge_Compiler_Scope.md","task":["TSK-CHER-002"]}
    rung
    EVIDENCED
    specifiedIn
    08_Knowledge_Repository/02_Knowledge_Compiler_Scope.md
  • checkStates
    {"packages validate":"PASS"}
    checks
    packages validate
    exercisedByTheRunner
    04_Canonical_Contracts/tooling/cher_validate.py
    id
    ENGINEERING_PLAN
    label
    Engineering plan / EEP
    note
    The canonical package, and the rules that refuse one before anything is built from it.
    producedBy
    {"doesNotCover":"anything about a running system. Every rule here reads a package document, so it can prove a package is internally honest and cannot prove that infrastructure is where the package says it is","module":"04_Canonical_Contracts/tooling/cher_validate.py"}
    ratio
    6 of 18 validation rules with an executable
    ratioTruthClass
    MEASURED
    reachesLiveSystem
    false
    refs
    {"check":["packages validate"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["04_Canonical_Contracts/tooling/cher_validate.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/13_Canonical_Engineering_Artefacts_Software_Factory.md","task":null}
    rung
    EXECUTED
    specifiedIn
    02_Target_Architecture/13_Canonical_Engineering_Artefacts_Software_Factory.md
  • checkStates
    {"security architecture specialist":"PASS"}
    checks
    security architecture specialist
    exercisedByTheRunner
    09_Factory_Bootstrap/plane-b/verify_specialist_security.py
    id
    SPECIALIST_ENGINEERING
    label
    Specialist engineering
    note
    Bounded specialists that produce the controls and requirements a package carries. TSK-CHER-004.
    producedBy
    {"doesNotCover":"seven of the target's eight roles. This is the second bounded role, Privacy and Regulatory Engineering. On the current corpus it derives nothing and reports why, which is a measurement of the knowledge repository rather than a failure of the role","module":"04_Canonical_Contracts/tooling/specialist_privacy.py"}, {"doesNotCover":"seven of the target's eight roles, and any use of a model. This is one bounded role, Security Architecture, deriving controls and their non-functional requirements from the controls the bound snapshot holds. It composes nothing: a requirement this specialist emits is a requirement a control already carried","module":"04_Canonical_Contracts/tooling/specialist_security.py"}, {"doesNotCover":"the other seven roles. This proves one bounded role derives from governed knowledge, cites what it derived from, stops at its decision limit and composes nothing","module":"09_Factory_Bootstrap/plane-b/verify_specialist_security.py"}
    qualifier
    PARTIAL
    ratio
    2 of 8 specialist roles implemented
    ratioTruthClass
    DECLARED
    reachesLiveSystem
    true
    refs
    {"check":["security architecture specialist"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["04_Canonical_Contracts/tooling/specialist_privacy.py","04_Canonical_Contracts/tooling/specialist_security.py","09_Factory_Bootstrap/plane-b/verify_specialist_security.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/12_Engineering_Agent_Specialist_Framework.md","task":["TSK-CHER-004"]}
    rung
    EVIDENCED
    specifiedIn
    02_Target_Architecture/12_Engineering_Agent_Specialist_Framework.md
  • checkStates
    {"VAL-CHER-AUTHORITY-001":"PASS"}
    checks
    VAL-CHER-AUTHORITY-001
    exercisedByTheRunner
    09_Factory_Bootstrap/plane-b/verify_decision_authority.py
    id
    GOVERNANCE_DECISION_RIGHTS
    label
    Governance / decision rights
    note
    Who may decide what, enforced rather than documented. D3 and above require a named human.
    producedBy
    {"doesNotCover":"the approval request path, which is the console's first write path and is not built; and decision recording by any application identity, since no application layer exists. What is enforced here is who may record an approval at all, attempted as five different identities","module":"09_Factory_Bootstrap/plane-b/verify_decision_authority.py"}
    reachesLiveSystem
    true
    refs
    {"check":["VAL-CHER-AUTHORITY-001"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["09_Factory_Bootstrap/plane-b/verify_decision_authority.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/08_Governance_Decision_Rights_Lifecycle.md","task":null}
    rung
    EVIDENCED
    specifiedIn
    02_Target_Architecture/08_Governance_Decision_Rights_Lifecycle.md
  • checkStates
    {"compiler is reproducible":"PASS"}
    checks
    compiler is reproducible
    exercisedByTheRunner
    09_Factory_Bootstrap/compiler/cher_compile_prompt.py
    id
    FACTORY_DISPATCH
    label
    Factory dispatch
    note
    The instruction leaves CHER through an adapter contract, not a vendor-shaped API.
    producedBy
    {"doesNotCover":"sending anything anywhere. This compiles the instruction and refuses to compile one that would be unsafe to send; a person carries it to the factory, and the adapter contract is the only factory-specific thing in the path","module":"09_Factory_Bootstrap/compiler/cher_compile_prompt.py"}
    reachesLiveSystem
    false
    refs
    {"check":["compiler is reproducible"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["09_Factory_Bootstrap/compiler/cher_compile_prompt.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/05_Integration_Layer.md","task":null}
    rung
    EXECUTED
    specifiedIn
    02_Target_Architecture/05_Integration_Layer.md
  • checkStates
    {"product projection":"PASS","result checker refuses what it should":"PASS"}
    checks
    result checker refuses what it should, product projection
    exercisedByTheRunner
    09_Factory_Bootstrap/compiler/check_execution_result.py, 09_Factory_Bootstrap/plane-b/verify_product_projection.py
    id
    RESULT_VALIDATION
    label
    Result validation
    note
    What came back, checked independently of whoever built it.
    producedBy
    {"doesNotCover":"the built system. This checks the result document a factory returned against the package it names, which is independent of the factory and is not an inspection of what it deployed","module":"09_Factory_Bootstrap/compiler/check_execution_result.py"}, {"doesNotCover":"anything the console does with this. It proves the read contract Plane A is given is honest about its own sources, counts and gaps; whether a screen renders them correctly is the slice's acceptance, checked by check_execution_result.py against the returned result","module":"09_Factory_Bootstrap/plane-b/verify_product_projection.py"}
    reachesLiveSystem
    true
    refs
    {"check":["result checker refuses what it should","product projection"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["09_Factory_Bootstrap/compiler/check_execution_result.py","09_Factory_Bootstrap/plane-b/verify_product_projection.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/13_Canonical_Engineering_Artefacts_Software_Factory.md","task":null}
    rung
    EVIDENCED
    specifiedIn
    02_Target_Architecture/13_Canonical_Engineering_Artefacts_Software_Factory.md
  • checkStates
    {"evidence ingestion path":"PASS","evidence store integrity":"PASS","packages validate":"PASS"}
    checks
    packages validate, evidence ingestion path, evidence store integrity
    exercisedByTheRunner
    04_Canonical_Contracts/tooling/cher_validate.py, 09_Factory_Bootstrap/plane-b/verify_evidence_ingestion.py, 09_Factory_Bootstrap/plane-b/verify_evidence_integrity.py
    id
    EVIDENCE_PROOF_CHAIN
    label
    Evidence / proof chain
    note
    Evidence kept so it can be produced later, and a chain from a claim back to its authority.
    producedBy
    {"doesNotCover":"the evidence store. SEM-TRACE-001 enforces the chain from a mandatory requirement back to an authority and forward to a validation, inside a package document. Nothing here persists evidence or produces it again later: cher.evidence_record is declared in TSK-CHER-001 and not built, so no claim in CHER can reach ASSURED","module":"04_Canonical_Contracts/tooling/cher_validate.py"}, {"doesNotCover":"assurance evaluation, which is what ASSURED means. This proves the loop from a check result to a durable record and back, and proves the status ladder will not advance a stage without one","module":"09_Factory_Bootstrap/plane-b/verify_evidence_ingestion.py"}, {"doesNotCover":"assurance evaluation. This proves evidence is persisted with its provenance and integrity and cannot be laundered into a higher truth class; it does not evaluate any claim against that evidence, which is what ASSURED means and what nothing in CHER does yet","module":"09_Factory_Bootstrap/plane-b/verify_evidence_integrity.py"}
    reachesLiveSystem
    true
    refs
    {"check":["packages validate","evidence ingestion path","evidence store integrity"],"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["04_Canonical_Contracts/tooling/cher_validate.py","09_Factory_Bootstrap/plane-b/verify_evidence_ingestion.py","09_Factory_Bootstrap/plane-b/verify_evidence_integrity.py"],"pack":null,"package":null,"relation":["cher.evidence_record"],"rule":null,"specification":"02_Target_Architecture/07_Evidence_Provenance_Traceability.md","task":["TSK-CHER-001"]}
    rung
    EVIDENCED
    specifiedIn
    02_Target_Architecture/07_Evidence_Provenance_Traceability.md
  • checkStates
    {}
    id
    ASSURANCE_CASE
    label
    Assurance case
    note
    Evidence assembled into an argument that the thing is fit, across the assurance modes.
    producedBy
    {"doesNotCover":"the twelfth lifecycle stage. Evaluates whether a scoped assurance claim is supported by admissible evidence over a governed basis, and refuses to elevate to ASSURED when it is not. It cannot grant assurance alone: a named human at the declared decision class is one of the conditions","module":"04_Canonical_Contracts/tooling/evaluate_assurance.py"}
    qualifier
    PARTIAL
    ratio
    1 of 7 assurance modes built
    ratioTruthClass
    DECLARED
    reachesLiveSystem
    false
    refs
    {"check":null,"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":["04_Canonical_Contracts/tooling/evaluate_assurance.py"],"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/06_Validation_Assurance_Architecture_Conformance.md","task":null}
    rung
    IMPLEMENTED
    specifiedIn
    02_Target_Architecture/06_Validation_Assurance_Architecture_Conformance.md
  • checkStates
    {}
    id
    RELEASE
    label
    Release
    note
    READY_FOR_RELEASE is a state in the target orchestrator. Nothing reaches it, because no orchestrator runs.
    qualifier
    NOT_STARTED
    ratio
    0 of 14 orchestrator states implemented
    ratioTruthClass
    DECLARED
    reachesLiveSystem
    false
    refs
    {"check":null,"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":null,"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/11_Integrated_Engineering_Orchestrator.md","task":null}
    rung
    SPECIFIED
    specifiedIn
    02_Target_Architecture/11_Integrated_Engineering_Orchestrator.md
  • checkStates
    {}
    id
    RUNTIME_CONTINUOUS_CONTROL
    label
    Runtime continuous control
    note
    Conformance after deployment, which is the claim that separates a control plane from a generator.
    qualifier
    NOT_STARTED
    reachesLiveSystem
    false
    refs
    {"check":null,"commit":"775ea10","correction":null,"decision":null,"evidenceArtefact":null,"executionResult":null,"gate":null,"module":null,"pack":null,"package":null,"relation":null,"rule":null,"specification":"02_Target_Architecture/09_Runtime_Operations_Continuous_Control.md","task":null}
    rung
    SPECIFIED
    specifiedIn
    02_Target_Architecture/09_Runtime_Operations_Continuous_Control.md

produced by python3 10_Archive_Integrity/build_status_dashboard.py

00b · Operating mode

A segment is reached when every stage it requires has been executed and is not partial.

  • partial
    false
    reached
    true
    requires
    CONTEXT_INFERENCE
    segment
    BUSINESS INTENT
    why
    a person describes an outcome and CHER derives the context
  • partial
    false
    reached
    true
    requires
    KNOWLEDGE_RESOLUTION
    segment
    CHER-DERIVED CONTEXT
    why
    the engineering basis is resolved from governed knowledge rather than written by hand
  • partial
    true
    reached
    false
    requires
    SPECIALIST_ENGINEERING
    segment
    SPECIALIST ENGINEERING
    why
    bounded specialists produce the controls the package carries
  • partial
    false
    reached
    true
    requires
    ENGINEERING_PLAN, GOVERNANCE_DECISION_RIGHTS
    segment
    CHER CONTROL PLANE
    why
    the package is validated and the decision rights are enforced
  • partial
    false
    reached
    true
    requires
    FACTORY_DISPATCH
    segment
    FACTORY
    why
    a factory builds it from a neutral instruction
  • partial
    false
    reached
    true
    requires
    RESULT_VALIDATION
    segment
    INDEPENDENT VALIDATION
    why
    what comes back is checked independently of whoever built it
  • partial
    true
    reached
    false
    requires
    EVIDENCE_PROOF_CHAIN, ASSURANCE_CASE
    segment
    ASSURANCE
    why
    the evidence becomes an assurance case rather than a set of passes
  • partial
    false
    reached
    false
    requires
    RELEASE, RUNTIME_CONTINUOUS_CONTROL
    segment
    RUNTIME CONTINUOUS CONTROL
    why
    conformance is maintained after deployment

produced by derived from panel 00

00c · Current build frontier

Every stage started and unfinished, plus the first not started. Not positioned by hand and deliberately not a single point.

  • firstNotStarted
    RELEASE
    frontier
    SPECIALIST_ENGINEERING, ASSURANCE_CASE, RELEASE
    startedAndUnfinished
    SPECIALIST_ENGINEERING, ASSURANCE_CASE

produced by derived from panel 00

01 · What you can do with it today

  • capability
    Author a package and have it independently validated
    evidence
    04_Canonical_Contracts/tooling/cher_validate.py
    how
    Two passes: JSON Schema, then semantic rules that schema cannot express - every reference resolves, every mandatory requirement traces to an authority and a validation, no agent resolves a D3 decision, truth classes do not launder, the geography gate matches the declared transfer basis.
  • capability
    Compile a factory-neutral instruction, reproducibly
    evidence
    09_Factory_Bootstrap/compiler/cher_compile_prompt.py
    how
    Byte-identical from unchanged inputs. Refuses a package that does not validate, a task naming a vendor, an unpassed pre-dispatch gate, an open D3 decision, and a capability the adapter does not declare.
  • capability
    Check what a factory returns, independently of the factory
    evidence
    09_Factory_Bootstrap/compiler/check_execution_result.py
    how
    Digest of the package it claims to have built, digest of the result itself, a disposition for every mandatory constraint, technologies against the contract, and prose that claims a control is met.
  • capability
    Stand up the governance plane in one command
    evidence
    09_Factory_Bootstrap/plane-b/bootstrap.sh
    how
    Migration and its 50 checks, tenant provisioning, the console boundary, the content-addressed snapshot, package load, then every runtime suite.
  • capability
    Prove the console cannot write governance, evidence or audit
    evidence
    09_Factory_Bootstrap/plane-b/verify_console_boundary.py
    how
    Connects as the console's own database role and attempts each write. The boundary is the revoked grant, not the absent feature.
  • capability
    Prove no machine identity can record a D3 or above
    evidence
    09_Factory_Bootstrap/plane-b/verify_decision_authority.py
    how
    Attempts a D5 approval as nobody, as a workload, as an agent, as an external system and as the platform administrator, and requires the named control to be what refuses each.
  • capability
    Turn ordinary business intent into a governed context profile
    evidence
    04_Canonical_Contracts/tooling/infer_context.py
    how
    Deterministic rules read one statement of intent plus any answers given. Every fact carries its derivation, its basis and a quotation from the request, no fact is OBSERVED, and a request that supports nothing produces no facts and a list of questions instead. Each question says why it is asked: a pack's applicability condition reads the fact, or the canonical contract requires it.
  • capability
    Say which issued packages a knowledge change reaches, and what their declared basis cannot support
    evidence
    08_Knowledge_Repository/tooling/assess_impact.py
    how
    Computes the impact set against two named snapshots, classifies the change from a closed set, and reports separately what a package's basis cannot support at all - a cited pack no snapshot has ever held, a rule defined off the basis, a governed identifier carrying a different check. It assigns no decision class and names no outcome: whether an affected version is re-issued, superseded or left alone is a governed decision under INV-CHER-007.
  • capability
    Say what changed between two governed snapshots
    evidence
    08_Knowledge_Repository/tooling/diff_snapshots.py
    how
    Membership in and out, and whether a supplied context resolves differently against each end. Deliberately not an impact assessment.
  • capability
    Resolve which knowledge applies to a context
    evidence
    08_Knowledge_Repository/tooling/resolve_applicability.py
    how
    Evaluates each pack's applicability conditions against the context facts and derives the binding, with what was rejected and why.
  • capability
    Prove the whole pack in one command
    evidence
    verify_pack.py
    how
    Every check, with a state each. A check that could not run is never reported as passed, and a declared failure names the correction that records it.

produced by each row names the module that makes it true

02 · What it cannot do yet

  • capability
    Run a package through its orchestrator states
    why
    TSK-CHER-005. Gates are evaluated by hand at the points a run would evaluate them. 0 of the target's 14 orchestrator states are implemented.
  • capability
    Produce an evidence pack
    why
    The store exists and holds observations with their provenance, integrity class and truth class. What does not exist is the assembled pack: the evidence for one package version, selected against its claims and exported as an artefact a third party can check.
  • capability
    Approve anything through the console
    why
    The approval request flow is the console's first write path and needs a grant change and a boundary-test change before it needs a screen.
  • capability
    Assess anything that already exists
    why
    6 of 7 assurance modes are deferred: brownfield, COTS, pre-acquisition, migration, portfolio, continuous assurance.
  • capability
    Test whether any of this improves what gets built
    why
    TSK-CHER-010, the Value Experiment. The compiler it was waiting for now exists, so what holds it is no longer a dependency: it needs a control arm, a blind human scorer and an external scanner, and none of that has run. No execution result exists for the task.

produced by designed, not built

03 · Coverage, measured

  • n
    6
    of
    18
    ratio
    validation rules with an executable
  • n
    3
    of
    3
    ratio
    knowledge compiler modules against declared behaviours
  • n
    3
    of
    6
    ratio
    console execution results held here
  • n
    0
    of
    10
    ratio
    engine tasks with a recorded execution result
  • n
    0
    of
    14
    ratio
    states
  • n
    2
    of
    8
    ratio
    roles
  • n
    1
    of
    7
    ratio
    modes

produced by python3 verify_pack.py --json --dsn <plane b>

04 · Every check, and its state

  • check
    manifest current
    lines
    216 artefacts, manifest and checksums current
    state
    PASS
  • check
    every artefact unchanged
    state
    PASS
  • check
    RFC 8785 canonicalisation
    lines
    canonical.py conforms
    state
    PASS
  • check
    package digests
    detail
    12 package versions stamped and correct
    state
    PASS
  • check
    packages validate
    detail
    20 conforming, donor v1.0.0 refused with its 6 expected findings, 16 expected knowledge-binding findings (C-45, C-46)
    state
    PASS
  • check
    console fixture current
    lines
    fixture current
    state
    PASS
  • check
    DDL invariants
    lines
    vocabularies match their producers; append-only claims are enforced
    state
    PASS
  • check
    architecture conformance
    detail
    C-39: ADR-CHER-003 and ADR-CHER-004 are unsigned, and the conformance check deliberately refuses to resolve an UNSIGNED heading (C-42). 001 and 002 are retired against the DEC records that govern. 003 needs a named D3 authority for the plane split; 004 needs D2 engineering review for placing readiness assessment in knowledge rather than in core. Both are decisions and neither is an edit. The unresolved identifiers are printed on the line below, so a third one appearing is visible against this entry rather than hidden by it.
    lines
    unresolved: ADR-CHER-003, ADR-CHER-004, ADR-CHER-005
    state
    KNOWN
  • check
    manifest counts
    detail
    8 stated numbers re-derived from the repository and they agree
    state
    PASS
  • check
    applicability resolution
    detail
    C-47: all ten hand-written knowledgeBindings differ from what resolves against their own context facts. Re-measured against the issued snapshot rather than the files (C-54), the finding stands at 10 of 10 and its content changed: the adapter-contract difference is gone, because the adapter is not a snapshot member, and two larger ones appeared - every package names DRAFT_CANDIDATE packs its own snapshot never held. This check runs the FILE corpus, which needs no database; the snapshot measurement is the authoritative one and runs under 'snapshot resolution'. Carries until a corrected version is issued: the issued versions are immutable and which is correct is a question for a person.
    lines
    issued versions are immutable, and which is correct is a question for a person.
    state
    KNOWN
  • check
    compiler is reproducible
    detail
    two compiles of TSK-CONSOLE-005 are byte-identical, 22612 bytes, compiledAt excluded
    state
    PASS
  • check
    result checker refuses what it should
    detail
    the good example is accepted with 0 blocking; the bad one is rejected with 10
    state
    PASS
  • check
    knowledge packs lint
    detail
    C-48: placement and the corpus metamodel version are properties of the cher-knowledge repository, and this pack is a distribution of the packs without that repository around them. Both are NOT RUN here and run in that repository's own CI. The pack contract check does run, against 04_Canonical_Contracts/schema/knowledge-pack.schema.json.
    lines
    [placement] NOT RUN: /home/claude/CHER_Delivery_Pack_v1.1/05_Knowledge_Packs is not the knowledge/ tree PLACEMENT.md describes, and placement is only meaningful against it, [metamodel] NOT RUN: the corpus metamodel version is not at /home/claude/CHER_Delivery_Pack_v1.1/08_Knowledge_Repository/repository-layout/metamodel/VERSION, so no pack's schemaVersion was compared against it. Pass --metamodel-version to name it.
    state
    KNOWN
  • check
    no em dashes
    detail
    the standing instruction, as a check rather than a habit
    state
    PASS
  • check
    Plane B constraint refusals
    lines
    12/12 checks passed
    state
    PASS
  • check
    VAL-CHER-AUTHORITY-001
    lines
    18/18 checks passed
    state
    PASS
  • check
    evidence store integrity
    lines
    23/23 checks passed
    state
    PASS
  • check
    evidence ingestion path
    lines
    19/19 checks passed
    state
    PASS
  • check
    snapshot resolution
    lines
    20/20 checks passed
    state
    PASS
  • check
    Plane A deployment record
    detail
    C-76: no projection carries the instant its values were taken. The product fixture reports 264 evidence rows in a store that holds 265, because verify_pack wrote an ingestion probe after the emit, and nothing on the projection lets a reader tell a stale screen from a wrong one. Open by intent rather than unrepaired: the fix changes both emitters and therefore both fixture digests, and re-emitting while Plane A is running the current set is C-70 exactly. It lands as contract 1.1.0 at the next dispatch, bundled so the supplied set moves once. Every other check on the record passes, including that the accepted result it descends from has not been edited.
    lines
    19/20 checks passed
    state
    KNOWN
  • check
    corpus diff
    lines
    26/26 checks passed (23 against real snapshots, 3 against constructed input this corpus cannot provide)
    state
    PASS
  • check
    context inference
    lines
    33/33 checks passed
    state
    PASS
  • check
    impact assessment
    lines
    33/33 checks passed (28 against real snapshots and issued packages, 5 against constructed input this corpus cannot provide)
    state
    PASS
  • check
    product projection
    lines
    20/20 checks passed
    state
    PASS
  • check
    security architecture specialist
    lines
    21/21 checks passed
    state
    PASS
  • check
    assurance evaluator
    lines
    32/32 checks passed
    state
    PASS
  • check
    privacy and regulatory specialist
    lines
    25/25 checks passed
    state
    PASS
  • check
    GATE-CONSOLE-PLANE-001
    lines
    17/17 checks passed
    state
    PASS

produced by python3 verify_pack.py --json --dsn <plane b>

05 · The engine: declared tasks

  • description
    Create the engine schema: tenants, identities, authorities, runs, packages, findings, evidence records, audit events and generic graph nodes and edges
    id
    TSK-CHER-001
  • description
    Implement the knowledge compiler: parse packs, validate against the pack schema, cross-check references, and issue content-addressed immutable snapsho
    id
    TSK-CHER-002
  • description
    Implement context inference producing one Solution Context Profile, with every material assumption recorded as a first-class object and every unmet ap
    id
    TSK-CHER-003
  • description
    Implement the bounded agent runtime with exactly two specialist roles, Security Architecture and Privacy/Regulatory Engineering, each with a declared
    id
    TSK-CHER-004
  • description
    Implement the five-state orchestrator with gate evaluation, running long operations on durable workers with retry, idempotency keys and a transactiona
    id
    TSK-CHER-005
  • description
    Implement the package compiler emitting an Engineering Execution Package that passes both schema validation and the semantic linter before it may leav
    id
    TSK-CHER-006
  • description
    Implement the factory adapter: compile factory-specific instructions from the canonical package, dispatch, capture the execution result and change man
    id
    TSK-CHER-007
  • description
    Implement the validator executing the eight rules against the canonical package and the change manifest, independently of the factory, and writing has
    id
    TSK-CHER-008
  • description
    Implement the operator console: runs, context and assumptions, packages, findings, evidence explorer, proof-chain view and approval requests
    id
    TSK-CHER-009
  • description
    Implement the experiment harness: one fixed brief, two arms of ten builds, arm-blind artefact export and a defect record per build
    id
    TSK-CHER-010

produced by 04_Canonical_Contracts/instances/EEP-CHER-ENGINE-001.v1.2.0.json

06 · The console: declared slices

  • description
    Implement the package list and the package view
    id
    TSK-CONSOLE-001
    resultHeldHere
    false
  • description
    Implement the findings view: every finding on a package, with its rule identifier, severity, the object it was raised against, and its narrative
    id
    TSK-CONSOLE-002
    resultHeldHere
    false
  • description
    Implement the proof-chain view
    id
    TSK-CONSOLE-003
    resultHeldHere
    false
  • description
    Implement the context and assumptions view
    id
    TSK-CONSOLE-004
    resultHeldHere
    true
  • description
    Move the gate display from the package's own gate
    id
    TSK-CONSOLE-005
    resultHeldHere
    true
  • description
    Turn the engineering console into the CHER product experience, on a second supplied fixture, without changing what the console is allowed to do
    id
    TSK-CONSOLE-006
    resultHeldHere
    true

produced by 09_Factory_Bootstrap/results/, against the tasks declared by EEP-CHER-CONSOLE-001.v1.6.0.json

07 · Validation rules and their executables

  • rule
    VAL-AUDIT-001
    state
    NOT RUN
  • rule
    VAL-AUTHZ-001
    state
    NOT RUN
  • rule
    VAL-CHER-AGENT-001
    state
    NOT RUN
  • executedBy
    {"doesNotCover":"auditing of refused attempts, which a database trigger cannot do because a refused write rolls back its own audit row, and which belongs to the unbuilt API layer the rule is scoped to; and the application identities in cher.actor, since no application layer exists to hold them","module":"09_Factory_Bootstrap/plane-b/verify_decision_authority.py"}
    rule
    VAL-CHER-AUTHORITY-001
    state
    EXECUTED
  • executedBy
    {"doesNotCover":"whether the engine, evidence store and audit log resolve to infrastructure in the home jurisdiction, which requires a deployment to inspect and nothing here inspects","module":"04_Canonical_Contracts/tooling/cher_validate.py"}
    rule
    VAL-CHER-GEO-001
    state
    EXECUTED
  • executedBy
    {"doesNotCover":null,"module":"09_Factory_Bootstrap/plane-b/verify_console_boundary.py"}
    rule
    VAL-CHER-PLANE-001
    state
    EXECUTED
  • rule
    VAL-CHER-SNAPSHOT-001
    state
    NOT RUN
  • rule
    VAL-CHER-TENANT-001
    state
    NOT RUN
  • rule
    VAL-CHER-TRACE-001
    state
    NOT RUN
  • executedBy
    {"doesNotCover":"factory results at rest, which check_execution_result.py covers at acceptance, and whether an evidence substrate can support the integrity class it is given, which nothing inspects","module":"04_Canonical_Contracts/tooling/cher_validate.py"}, {"doesNotCover":"package-side provenance and evidence contracts, which cher_validate.py covers, and whether a substrate supports its claimed integrity class, which nothing inspects","module":"09_Factory_Bootstrap/compiler/check_execution_result.py"}, {"doesNotCover":"the package-side provenance and evidence contracts, which cher_validate.py covers. This covers the ingestion path: what may and may not be written when a check result becomes a record","module":"09_Factory_Bootstrap/plane-b/verify_evidence_ingestion.py"}, {"doesNotCover":"package-side provenance and evidence contracts, which cher_validate.py covers, and factory results at acceptance, which check_execution_result.py covers. This covers the stored path only: what the evidence store will and will not persist","module":"09_Factory_Bootstrap/plane-b/verify_evidence_integrity.py"}
    rule
    VAL-CHER-TRUTH-001
    state
    EXECUTED
  • rule
    VAL-CHER-WORKER-001
    state
    NOT RUN
  • rule
    VAL-PERF-001
    state
    NOT RUN
  • rule
    VAL-PURPOSE-001
    state
    NOT RUN
  • rule
    VAL-RETENTION-001
    state
    NOT RUN
  • rule
    VAL-SECRETS-001
    state
    NOT RUN
  • rule
    VAL-TENANT-001
    state
    NOT RUN
  • executedBy
    {"doesNotCover":"whether factory output conforms to the package, which check_execution_result.py covers for a submitted result","module":"04_Canonical_Contracts/tooling/cher_validate.py"}, {"doesNotCover":"requirement-level upstream and downstream links, which cher_validate.py covers on the package","module":"09_Factory_Bootstrap/compiler/check_execution_result.py"}
    rule
    VAL-TRACE-001
    state
    EXECUTED
  • executedBy
    {"doesNotCover":"whether a recorded ground is supported by evidence, which this module does not fetch; the rule is declared MANUAL_REVIEW and the automated part of it is the declaration check only","module":"04_Canonical_Contracts/tooling/cher_validate.py"}
    rule
    VAL-XBORDER-001
    state
    EXECUTED

produced by module-level EXECUTES declarations

08 · What remains, in dependency order

  • n
    1
    title
    Three engine tasks are substantially built and CHER does not know it
  • n
    2
    title
    The three accepted results cite unresolvable digests
  • n
    3
    title
    Narrow the console read surface
  • n
    4
    title
    ADR-CHER-003 and ADR-CHER-004 are unsigned
  • n
    5
    title
    Twelve of eighteen validation rules have no executable
  • n
    6
    title
    The fixture has moved and the console has not been given it
  • n
    7
    title
    A factory's assumptions reach no reader but the acceptance report
  • n
    8
    title
    sourceControlRef is available and unpopulated
  • n
    9
    title
    Schema 2.0.0
  • n
    10
    title
    A dispatched prompt is not reproducible across a compiler change
  • n
    11
    title
    What the impact assessment found, and cannot close
  • n
    12
    title
    Three of the five accepted console results are not in the archive
  • n
    13
    title
    C-54 is a governance decision, not an edit

produced by HANDOVER.md section 4

09 · Governance plane

  • policies
    24
    tables
    20
    triggers
    31

produced by counted from the DDL in 08_Knowledge_Repository/ddl and 09_Factory_Bootstrap/plane-b

10 · Corrections

  • bySeverity
    {"Critical":9,"High":35,"Low":8,"Medium":29}
    entries
    81
    provenanceFrom
    C-31
    statingHowTheyWereFound
    41
    total
    81

produced by 07_Corrections_Register/01_Defect_Corrections_from_v1.0.md

11 · Knowledge compiler

  • behavioursDeclared
    3
    bindings
    12
    bindingsThatDisagree
    12
    modules
    3

produced by python3 08_Knowledge_Repository/tooling/resolve_applicability.py --all --json