Role Capability Map Review¶
The role capability map is a versioned control-plane contract artifact. It binds manifest repository classes to role groups, role groups to verifier capability profiles, and role groups to graph-permission rules.
This document records the review questions used to keep
data/schema/role-capability-map.toml durable.
It is intentionally descriptive rather than generated.
The generated checks should validate the map;
this document explains what the map is expected to mean.
Review Status¶
Current status: draft
Release target: v1.0.0 control-plane consistency
Primary artifact:
Companion schema:
Design specification:
Review Principle¶
The map should encode durable authority boundaries. The goal is to make drift inspectable, diagnosable, and testable. The same class and role information is consumed by:
- manifest graph verification
- Accountable Record stage profile derivation
- future Rust conformance checks
Review Dimensions¶
Each role group should be reviewed across four dimensions:
- Role identity
- Stage capability profile
- Graph authority permissions
- Release status
These dimensions are separate. A repository can be active but not release-gate authoritative. A repository can be consumed as tooling without becoming semantic authority. A role group can emit reports without exporting contract artifacts.
Role Group Review Table¶
Use this table to review each role group declared in
data/schema/role-capability-map.toml.
| Role group | Manifest classes | Validates target materials | Validates resolution | Requires lock artifact | Exports contract artifacts | Emits human reports | Layer order reviewed | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|
| adapter | adapter | false | false | false | false | false | no | draft | Review whether adapters should emit reports. |
| administration | admin | false | false | false | false | true | no | draft | Administrative repos may coordinate checks without becoming semantic authority. |
| constitution | constitution | true | true | true | false | true | no | draft | Review whether constitution exports contract artifacts or remains governance-only. |
| contract | contract | true | true | true | true | true | no | draft | Contract repos are semantic contract surfaces. |
| data | data_companion | true | true | true | false | true | no | draft | Review whether all data companion repos require locks. |
| documentation | docs | false | false | false | false | true | no | draft | Documentation is curated output, not semantic authority. |
| formal_contract | formal_contract | true | true | true | true | true | no | draft | Bridges formal theory to operational contract artifacts. |
| govsrc | govsrc | true | true | true | false | true | no | draft | Source registries can be terminal entry points. |
| implementation | implementation | true | true | true | false | true | no | draft | Implementations verify or consume contracts; they do not define contract semantics. |
| kernel | kernel | true | true | true | true | true | no | draft | Review relation between kernel and constitution. |
| manifest_schema | manifest_schema | true | true | true | true | true | no | draft | Owns manifest structure and graph-validation semantics. |
| mapping | mapping | true | true | true | false | true | no | draft | Mapping should not leak interpretation into substrate or boundary layers. |
| mapspec | mapspec | true | true | true | true | true | no | draft | Mapping specifications may export structured artifacts. |
| pilot | pilot | false | false | false | false | true | no | draft | Pilots may remain incomplete without blocking v1.0.0. |
| policy | policy | true | true | true | false | true | no | draft | Review whether policy should export contract-like artifacts. |
| profile | profile | true | true | true | false | true | no | draft | Profiles constrain application contexts but should not override substrate. |
| record_system | record_system | true | true | true | false | true | no | draft | Domain record systems should consume contracts, not define foundation semantics. |
| regimes | regimes | true | true | true | false | true | no | draft | Review whether regimes should be folded into contract artifacts later. |
| rules | rules | true | true | true | false | true | no | draft | Jurisdictional or context rules remain external to substrate. |
| schema | schema | true | true | true | true | true | no | draft | Domain or exchange schemas may export machine-readable artifacts. |
| specification | specification | true | true | true | false | true | no | draft | Specifications may be provenance, transitional, or canonical depending on manifest. |
| theory | theory | true | true | true | false | true | no | draft | Theory exports formalized knowledge, not operational machine-readable artifacts. |
| tooling | engine | false | false | false | false | true | no | draft | Tooling dependencies are real dependencies but not semantic authority edges. |
| verification | verification | true | true | true | false | true | no | draft | Verification repos check records without defining contract authority. |
| verifier | verifier | true | true | true | false | true | no | draft | Standalone verifiers consume contracts and emit reports. |
Capability Questions¶
For each role group, answer these questions before marking it reviewed.
validates_target_materials¶
Does this role group validate authored or declared target materials?
Examples of target materials include manifests, source declarations, element declarations, catalog declarations, theory surfaces, mapping tables, schemas, rules, source registries, and record inputs.
Use true only when the role group is expected to inspect domain or contract
materials as part of its own verification profile.
validates_resolution¶
Does this role group validate references, dependency resolution, provenance links, or lock state?
Use true when the role group participates in reference resolution or must prove
that its upstream references are coherent.
requires_lock_artifact¶
Does this role group require a lock, digest, or provenance artifact even when the artifact is empty-but-valid?
Use true when absence of external references should still be represented as an
empty-but-valid lock state rather than omitted.
exports_contract_artifacts¶
Does this role group export machine-readable artifacts consumed downstream as contract or schema authority?
Use true for contract, formal contract, schema, kernel, manifest schema, or
other roles that intentionally provide machine-readable authority surfaces.
Use false for roles that report, verify, document, or implement without
becoming contract authority.
emits_human_reports¶
Does this role group emit human-readable reports as part of its expected verification or governance surface?
Use true when a verifier, checker, or release gate should produce Markdown or
other human-readable review material.
Graph-Permission Review¶
Graph permissions apply to semantic dependency edges only. Tooling dependencies are still real dependencies. They must resolve and they must name valid repositories. They do not confer semantic authority. Review each graph rule separately.
SI01: Required semantic dependency graph is acyclic¶
Question:
Expected answer: no
Review status: draft
SI02: All declared dependencies resolve¶
Question:
Does every declared dependency, regardless of kind, resolve to a real repository
with a valid manifest?
Expected answer: yes
Review status: draft
SI03: Provides are real¶
Question:
Expected answer: yes
Review status: draft
SI04: Class registry is satisfied¶
Question:
Expected answer: yes
Review status: draft
SG01: No core depends on domain¶
Question:
Can a core, substrate, boundary, kernel, or manifest-schema role depend
semantically on domain, application, example, govsrc, or record-system roles?
Expected answer: no
Review status: draft
SG02: Interpretation stays external¶
Question:
Expected answer: no
Review status: draft
SG03: Implementations are not semantic authority¶
Question:
Can contract, formal-contract, schema, or record-system roles depend
semantically on implementation, verification, or verifier roles?
Expected answer: no
Review status: draft
SG04: Operational systems do not consume theory directly¶
Question:
Can operational implementations, record systems, applications, or verifiers
depend semantically on theory repos directly?
Expected answer: no. The semantic path should run through formal-contract or other declared contract surfaces.
Review status: draft
SG05: Layer monotonicity¶
Question:
Expected answer: yes
Review status: draft
CG01: No non-terminal orphans¶
Question:
Is every active non-terminal role group consumed by at least one other repo,
unless it declares an accepted status or is an entry-point root?
Expected answer: yes
Review status: draft
CG02: Formalization sources are classified¶
Question:
Is each formalization result classified per artifact as canonical, auxiliary,
transitional, or superseded?
Expected answer: yes
Review status: draft
Surface bucket review¶
Theory surface mirrors must expose these semantic-role buckets:
A bucket may be empty, but it must exist.
Review question:
Does each theory repo classify public Lean symbols by semantic role rather than
Lean keyword category?
Expected answer: yes
Review status: draft
Formalization status review¶
Formalization status belongs under the provided artifact, not at repo level.
Example:
[provides.formalization.identity_regime_lower_bound]
kind = "lean-theorem-set"
status = "canonical"
symbols = [
"IdentityRegime.lowerBound",
]
Allowed statuses:
Review question:
Does each formalization artifact declare whether it is canonical, auxiliary,
transitional, or superseded?
Expected answer: yes
Review status: draft
Dependency kind review¶
Each dependency should declare a kind.
First required kinds:
Reserved kinds:
Review question:
Expected answer: yes
Review status: draft
Release-Gate Checklist¶
Before v1.0.0, complete these checks:
- Every manifest class in
manifest-schema.tomlappears inrole_groups. - Every role group in
role_groupshas one capability profile. - Every role group in
role_groupshas one layer-order value. - Every graph rule has a stable
SE.ORG.*diagnostic. - Every required theory surface bucket is present in each theory mirror.
- Every semantic dependency kind is reviewed.
- Every tooling dependency kind is reviewed.
- Every canonical formalization source is declared per provided artifact.
- Every duplicate formalization source is classified.
- Every accepted orphan has a recorded reason.
- The AR verifier consumes
role_groupsandcapability_profiles. - The graph verifier consumes
role_groupsandgraph_permissions. - Future Rust conformance expectations reference this same artifact.
Change policy¶
Changes to data/schema/role-capability-map.toml should be treated as
contract changes, not ordinary configuration edits.
Every change should answer:
Which consumer is affected?
Which capability, graph rule, or role binding changes?
Is this a breaking change for any verifier?
Does the change require a fixture update?
Does the change require a diagnostic-code update?
Open review items¶
- Confirm whether
adaptershould emit human reports. - Confirm whether
constitutionshould export contract artifacts. - Confirm whether all
dataroles require lock artifacts. - Confirm relation between
kernelandconstitution. - Confirm whether
policyshould export contract-like artifacts. - Confirm whether
regimesshould eventually feed formal-contract exports. - Confirm layer order values after the first graph report is generated.
- Confirm the first canonical formalization declarations under
[provides.formalization.*].