Complete REUSE-WP-0003: registry CLI, docs alignment, and coverage

Align INTENT.md with delivered layout, add CapabilityRegistryConcept guide,
extend schema with promotion_history, ship reuse-surface validate/query/export
CLI, register three more helix_forge capabilities, and refresh SCOPE and gap
analysis to reflect A3 tooling and D5/A3/C4/R2 self-assessment.
This commit is contained in:
2026-06-15 01:12:09 +02:00
parent 80bccc4bbb
commit 0dbef6d1a3
20 changed files with 1051 additions and 165 deletions

View File

@@ -0,0 +1,111 @@
---
id: capability.feature-control.rollout
name: Feature Rollout Control
summary: Gradually expose features to subjects across tenants, domains, groups, or cohorts using rollout rules and staged availability.
owner: feature-control
status: draft
domain: helix_forge
tags:
- feature-control
- rollout
- planning
maturity:
discovery:
current: D4
target: D6
confidence: medium
rationale: >
Rollout is a distinct bounded behavior from single-point evaluation, with
research references in the feature-control domain and maturity standard.
availability:
current: A2
target: A4
confidence: low
rationale: >
Rollout logic may exist in source modules but is not yet consistently
exposed as a standalone SDK or API surface distinct from evaluate.
external_evidence:
completeness:
level: C2
name: Partial
confidence: low
basis: scope_vs_intent_and_consumer_expectations
satisfied_expectations:
- rollout concepts documented adjacent to feature evaluation
broken_expectations:
- no dedicated rollout artifact called out separately from evaluate
- percentage and cohort rollout variants not indexed independently
out_of_scope_expectations:
- billing-driven entitlements
reliability:
level: R1
name: Fragile
confidence: low
basis: consumer_quality_signals
known_reliability_risks:
- rollout behavior may be conflated with evaluate in consumer mental models
discovery:
intent: >
Control how features become available over time and across cohorts without
conflating rollout policy with authorization or billing.
includes:
- staged rollout rules
- cohort and context targeting for rollout
- explainable rollout state
excludes:
- one-time feature evaluation only
- authorization decisions
- billing entitlements
assumptions:
- feature evaluation capability exists for final availability decisions
use_cases:
- ucc.feature-control.domain-rollout
research_memos:
- specs/CapabilityMaturityStandard.md
availability:
current_level: A2
target_level: A4
current_artifacts:
- feature-control/packages/feature-control-sdk
target_artifacts:
- feature-control/services/feature-control-api
consumption_modes:
- source module
- SDK
relations:
depends_on:
- capability.feature-control.evaluate
supports: []
related_to:
- capability.feature-control.visibility
evidence:
documentation:
- specs/CapabilityMaturityStandard.md
tests: []
consumer_feedback: []
bug_reports: []
incidents: []
consumer_guidance:
recommended_for:
- planning staged feature exposure separate from binary evaluation
not_recommended_for:
- simple on/off evaluation without rollout semantics
- entitlement or billing ownership
known_limitations:
- distinguish carefully from capability.feature-control.evaluate
---
# Feature Rollout Control
## Overview
Rollout governs how availability changes over time and across cohorts. It
complements evaluation, which answers whether a feature is available for a
subject in a context right now.

View File

@@ -0,0 +1,115 @@
---
id: capability.identity.subject-resolution
name: Identity Subject Resolution
summary: Resolve who or what is acting in a context by mapping principals, accounts, actors, and identifiers to a stable subject model.
owner: identity-canon
status: draft
domain: helix_forge
tags:
- identity
- subject
- architecture
maturity:
discovery:
current: D3
target: D5
confidence: medium
rationale: >
Subject/principal terminology is explored in identity-canon conflict maps
and conceptual model, but dedicated use-case grounding is incomplete.
availability:
current: A0
target: A4
confidence: low
rationale: >
Canon and research artifacts exist; no standalone resolver service or SDK
is registered yet.
external_evidence:
completeness:
level: C1
name: Fragmentary
confidence: low
basis: scope_vs_intent_and_consumer_expectations
satisfied_expectations:
- overloaded subject and principal terms are mapped as candidates
broken_expectations:
- no runtime resolver artifact
- canonical subject model not finalized across all actor types
out_of_scope_expectations:
- authentication protocol implementation
- credential storage
reliability:
level: R0
confidence: low
basis: consumer_quality_signals
known_reliability_risks:
- draft terminology may change during source-note backfill
discovery:
intent: >
Give planners and implementers a consistent subject concept for authorization,
feature control, tenancy, and agent workflows without collapsing product-specific
identity models.
includes:
- subject vs principal vs account distinctions
- actor type modeling
- identifier resolution concepts
excludes:
- authentication execution
- credential issuance
- directory provisioning
assumptions:
- vocabulary canonicalization supports but does not replace subject resolution
use_cases:
- UC-RS-004
research_memos:
- identity-canon/terminology/TerminologyConflictMap.md
- identity-canon/model/ConceptualModel.md
availability:
current_level: A0
target_level: A4
current_artifacts:
- identity-canon/model/ConceptualModel.md
- identity-canon/canon/CanonicalGlossary.md
target_artifacts:
- identity-canon/packages/subject-resolution-sdk
consumption_modes:
- informational
relations:
depends_on:
- capability.identity.vocabulary-canonicalize
supports:
- capability.feature-control.evaluate
- capability.statehub.workstream-coordinate
related_to: []
evidence:
documentation:
- identity-canon/canon/CanonicalGlossary.md
- identity-canon/scenarios/ScenarioTests.md
tests: []
consumer_feedback: []
bug_reports: []
incidents: []
consumer_guidance:
recommended_for:
- architecture planning where subject/principal/account terms overlap
not_recommended_for:
- runtime authentication or token validation
- treating draft canon as finalized resolver behavior
known_limitations:
- resolver artifacts are not yet available
---
# Identity Subject Resolution
## Overview
Subject resolution defines how actors and identifiers map to a stable subject
concept for downstream capabilities such as feature evaluation and coordination.
Today this capability is planning-heavy (D3/A0).

View File

@@ -20,34 +20,38 @@ maturity:
registry model is bounded in INTENT.md, but concrete promotion workflows
are not yet grounded in registry artifacts.
availability:
current: A0
current: A3
target: A3
confidence: medium
rationale: >
Registration is currently manual through Markdown entries and the index.
A CLI validator/query tool would raise availability to A3.
Registration and maintenance are automatable through the reuse-surface CLI
for validate, query, and export workflows.
external_evidence:
completeness:
level: C1
name: Fragmentary
confidence: low
level: C2
name: Partial
confidence: medium
basis: scope_vs_intent_and_consumer_expectations
satisfied_expectations:
- capability IDs can be assigned in registry entries
- maturity vectors can be recorded at registration time
- CLI validate, query, and export support registry workflows
- promotion history can be recorded in entries
broken_expectations:
- no automated duplicate detection yet
- no promotion history tracking yet
out_of_scope_expectations:
- hosting registered capabilities
- enforcing implementation architecture
reliability:
level: R0
confidence: low
level: R2
name: Tolerable
confidence: medium
basis: consumer_quality_signals
known_reliability_risks:
- registry consistency depends on manual index maintenance
- index drift still possible if authors skip validate
- CLI requires local venv install
- schema ID pattern required a fix during WP-0003 dogfooding
discovery:
intent: >
@@ -72,17 +76,17 @@ discovery:
- specs/ProductRequirementsDocument.md
availability:
current_level: A0
current_level: A3
target_level: A3
current_artifacts:
- templates/capability-entry.template.md
- registry/indexes/capabilities.yaml
target_artifacts:
- tools/validate-registry
- tools/query-registry
- reuse_surface/cli.py
target_artifacts: []
consumption_modes:
- informational
- markdown authoring
- cli
relations:
depends_on: []
@@ -111,6 +115,26 @@ consumer_guidance:
known_limitations:
- manual index updates are required after adding an entry
- duplicate detection is guidance-only in the MVP
promotion_history:
- date: "2026-06-14"
dimension: discovery
from: D0
to: D1
rationale: Added name, summary, and registry-first intent from INTENT.md.
author: reuse-surface
- date: "2026-06-15"
dimension: discovery
from: D1
to: D3
rationale: Bounded scope with UC-RS-001 and MVP registry artifacts in specs/.
author: reuse-surface
- date: "2026-06-15"
dimension: availability
from: A0
to: A3
rationale: reuse-surface CLI shipped validate, query, and export commands.
author: reuse-surface
---
# Capability Registration

View File

@@ -0,0 +1,117 @@
---
id: capability.statehub.workstream-coordinate
name: Workstream And Task Coordination
summary: Track active workstreams, tasks, progress, and consistency across domain repositories through a local-first coordination service.
owner: state-hub
status: draft
domain: helix_forge
tags:
- state-hub
- coordination
- workplans
maturity:
discovery:
current: D4
target: D5
confidence: medium
rationale: >
State Hub intent, scope, and ADR-001 workplan conventions are documented;
concrete agent workflows are still being standardized across repos.
availability:
current: A4
target: A6
confidence: medium
rationale: >
Consumers integrate through the State Hub HTTP API and custodian scripts.
Managed platform operation is the natural target for multi-repo agents.
external_evidence:
completeness:
level: C3
name: Functional Core
confidence: medium
basis: scope_vs_intent_and_consumer_expectations
satisfied_expectations:
- workstream and task tracking via HTTP API
- workplan file to database consistency sync
- progress event logging
broken_expectations:
- no native MCP server for all agent environments
out_of_scope_expectations:
- replacing git-backed workplan canon
- identity authority
reliability:
level: R2
name: Tolerable
confidence: low
basis: consumer_quality_signals
known_reliability_risks:
- depends on local hub availability and operator maintenance
discovery:
intent: >
Give agents and operators a live, queryable memory of work across domains
without replacing file-backed workplans and governance canon.
includes:
- workstream and task indexing
- progress and decision events
- repo consistency synchronization
- inbox and human-review flags
excludes:
- canon ownership
- task factory for all work origins
- identity authority
assumptions:
- workplans remain authoritative in git for ADR-001 repos
use_cases:
- UC-RS-013
research_memos:
- state-hub/INTENT.md
availability:
current_level: A4
target_level: A6
current_artifacts:
- state-hub/api/
- state-hub/scripts/consistency_check.py
target_artifacts:
- state-hub/platform/state-hub-service
consumption_modes:
- service API
- HTTP REST
relations:
depends_on: []
supports:
- capability.registry.register
related_to:
- capability.statehub.progress-log
evidence:
documentation:
- state-hub/INTENT.md
- state-hub/AGENTS.md
tests: []
consumer_feedback: []
bug_reports: []
incidents: []
consumer_guidance:
recommended_for:
- cross-repo agent orientation and task status sync
- logging progress after workplan changes
not_recommended_for:
- storing canonical requirements or architecture canon
- assuming hub state overrides git workplan files
known_limitations:
- requires running State Hub locally or via tunnel
---
# Workstream And Task Coordination
## Overview
State Hub provides live coordination for helix_forge repositories while
workplans remain file-backed. This entry represents the service/API consumption
mode for cross-repo agent work.