Retire completed roadmaps and collect future work

This commit is contained in:
2026-11-04 16:04:32 +00:00
parent 260256d722
commit bc1d1d5844
12 changed files with 80 additions and 2197 deletions

View File

@@ -43,7 +43,7 @@ ckprompt v0.2.0; the published v0.2.0 pair remains matched. See the
- [Maintainer release procedure](docs/release.md)
- [Architecture policy](docs/policy/architecture.md)
- [Policy provenance](docs/policy/meta.md)
- [Deferred capabilities and proposals](docs/roadmap/deferred.md)
- [Future work and proposals](docs/roadmap/future.md)
CommandKit is licensed under the [LGPL v3.0 or later](COPYRIGHT).

View File

@@ -15,8 +15,8 @@ canonical documentation owners.
For application-owned aliases, use the
[CLI reference](cli.md#application-owned-aliases),
[Go declarations and GoDoc](../command.go), and
[ADR 0013](adr/0013-application-alias-routing.md). Keep deferred capabilities in
[deferred work](roadmap/deferred.md).
[ADR 0013](adr/0013-application-alias-routing.md). Keep unfinished work and
capability proposals in [future work](roadmap/future.md).
## Initial orientation
@@ -54,17 +54,17 @@ For application-owned aliases, use the
| Task | Read |
| --- | --- |
| Implemented CLI behavior | [CLI reference](cli.md) |
| Invocation reporting | [CLI reporting contract](cli.md#reporting), [consumer guide](consumers/guide.md#invocation-reporting), core GoDoc, [reporting roadmap](roadmap/reporting.md), and [ADR 0014](adr/0014-invocation-reporting.md) |
| Invocation reporting | [CLI reporting contract](cli.md#reporting), [consumer guide](consumers/guide.md#invocation-reporting), core GoDoc, and [ADR 0014](adr/0014-invocation-reporting.md) |
| Application-owned aliases | [CLI reference](cli.md#application-owned-aliases), [Go declarations and GoDoc](../command.go), [consumer guide](consumers/guide.md#application-owned-aliases), and [ADR 0013](adr/0013-application-alias-routing.md) |
| Public API usage and standard-suite wiring | [Consumer guide](consumers/guide.md), [ckconfig package](../integrations/ckconfig/doc.go), [ckprompt package](../integrations/ckprompt/doc.go), and applicable ADRs |
| Dispatcher, registration, or typed inputs | [Command declarations](../command.go), [bindings](../binding.go), [invocation](../execute.go), and [ADR 0001](adr/0001-command-dispatch-and-inputs.md), refined by [ADR 0011](adr/0011-captured-declarations-and-installation.md) |
| Backend integration | [Internal overview](internal/overview.md), [adapter implementation](../internal/cobra/parse.go), and [ADR 0002](adr/0002-cobra-backend.md) |
| Public facade or adapter boundaries | [Internal overview](internal/overview.md), [ADR 0003](adr/0003-public-facade.md) as scoped by [ADR 0010](adr/0010-optional-integration-modules.md), and the architecture policy |
| Module ownership and public contracts | [Consumer guide](consumers/guide.md), [internal overview](internal/overview.md), [core GoDoc](../doc.go), [ckconfig GoDoc](../integrations/ckconfig/doc.go), [ckprompt GoDoc](../integrations/ckprompt/doc.go), and applicable ADRs |
| Deferred capabilities and proposals | [Deferred work](roadmap/deferred.md) |
| Unfinished work and capability proposals | [Future work](roadmap/future.md) |
| Completed planning and implementation history | [Completed dependency upgrades, PromptKit validation, discovery, smoke testing, and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/a6278a6d9f940948ec7745955d745523aaff0734/docs/roadmap), [earlier ConfigKit and PromptKit adoption and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/8a82620dfcd34c0ae56bbecb0d014eb49786bc92/docs/roadmap), [earlier retired records](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/6bc37e7ba64ca0920b4d0cc77039a18d4006b24e/docs/roadmap), and [completed module implementation and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/bb2528532b78d30e52f5082e491a5e1fd9b0fb99/docs/roadmap) in Git history |
| Release tooling | [Maintainer release procedure](release.md) |
| Performance maintenance | [Internal overview](internal/overview.md), existing benchmarks, and [deferred efficiency work](roadmap/deferred.md#possible-further-efficiency-work) |
| Performance maintenance | [Internal overview](internal/overview.md), existing benchmarks, and [possible further efficiency work](roadmap/future.md#possible-further-efficiency-work) |
| Licensing or incorporated material | [Project grant](../COPYRIGHT), [license texts](../LICENSE), and [attribution](oss/attribution.md) |
| Policy provenance | [Policy provenance](policy/meta.md) |
@@ -72,11 +72,12 @@ Read all applicable rows for cross-cutting work. Go declarations and GoDoc own
current exported contracts; the CLI reference owns operator behavior. The
[maintainer release procedure](release.md) owns current release operations.
Completed roadmaps and implementation evidence remain available in Git history,
including [application-owned alias planning, implementation, and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/f42aefe02d84e154855ac922173e6f367180929f/docs/roadmap)
including [ConfigKit adoption, reporting, release-tooling, and remediation records](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/260256d722f52bc947103d078cb67a129f4cc9f1/docs/roadmap),
[application-owned alias planning, implementation, and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/f42aefe02d84e154855ac922173e6f367180929f/docs/roadmap)
and [PromptKit requirements implementation and remediation](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/40499a43608bf60d1104cd8ee24caf18ac0e9e66/docs/roadmap).
These records are not assignments to restart completed work. The
[deferred-work record](roadmap/deferred.md) preserves future capability targets
and unselected proposals separately from current contracts.
[future work record](roadmap/future.md) preserves unfinished work and unselected
proposals separately from current contracts.
## Documentation validation

View File

@@ -56,7 +56,7 @@ invariant, not from the convenience of the selected implementation. Explain the
specific failure it prevents and why a less restrictive implementation cannot
satisfy the contract.
The [deferred-work record](../roadmap/deferred.md) distinguishes future proposals
The [future work record](../roadmap/future.md) distinguishes future proposals
from accepted contracts. Do not approve proposed commands, grammar, statuses, or
illustrative API declarations
merely by repeating them in implementation or tests. Cobra/pflag defaults do not

View File

@@ -294,7 +294,7 @@ recommended for application-owned commands; their names and domain behavior
remain application choices.
Accepted ADRs and current CLI/API owners define selected contracts; the
[deferred-work record](../roadmap/deferred.md) preserves remaining proposals.
[future work record](../roadmap/future.md) preserves remaining proposals.
Backend defaults and incidental consumer behavior do not establish new contracts.
A replacement backend must preserve published CLI behavior or follow deliberate compatibility
changes under the release policy.

View File

@@ -1,247 +0,0 @@
# ConfigKit v1.1.0 adoption
Status: the selected ckconfig adoption is implemented and prepared for a later
release; source publication and ckprompt migration remain pending. This roadmap
defines the embedded `ckconfig` dependency and source-report target.
[Selected scope decisions](#selected-scope-decisions) distinguish adoption from
deferred extensions. The [implementation plan](implementation.md) owns stage
order, concrete changes, and validation steps.
## Purpose and authority
Adopt `go.maximumdirect.net/configkit v1.1.0` and expose its capabilities through
CommandKit's existing configuration commands in a neutral, consistent way.
ConfigKit owns configuration loading, composition, resolution, inspection, and
their semantics. CommandKit owns command selection, coordination, presentation,
and invocation outcomes. Adapt upstream capabilities faithfully, using current
upstream terminology, without adding independent guarantees or commentary
qualifying upstream limitations.
This follows the clarified integration goal and the existing
[architecture policy](../policy/architecture.md#backend-and-library-integration).
The [agent](../policy/agents.md), [testing](../policy/testing.md),
[documentation](../policy/documentation.md), [release](../policy/release.md), and
[OSS](../policy/oss.md) policies govern implementation. This roadmap records
the selected target; current contracts remain in GoDoc, the
[CLI reference](../cli.md), and the [consumer guide](../consumers/guide.md) until
implementation lands.
## Current baseline and review evidence
Initial review revision: `60841227b7e2928d3fefd3326f0e3f4694f41e6d`.
API coverage follow-up revision: `7a9e3a1f7cdd0475f06a63ce0aababefe08651bc`;
library code and module selections are unchanged between these revisions.
The completed [dependency alignment roadmap](integration-dependency-alignment.md)
records the preceding vanity migration and releases.
| Component | Current selection |
| --- | --- |
| Core | `go.maximumdirect.net/commandkit v0.3.0`; Go 1.26.0 minimum |
| ckconfig | `go.maximumdirect.net/ckconfig v0.2.0`; Go 1.26.0 minimum |
| ckconfig dependencies | Core v0.3.0, `gitea.maximumdirect.net/eric/ConfigKit v0.8.0`, YAML v4.0.0-rc.6 |
| ckprompt | v0.2.0; ckconfig v0.2.0, old-path ConfigKit v0.8.0 and PromptKit v0.16.0; Go 1.26.0 minimum |
| Maintainer release toolchain | Go 1.26.8, with `GOTOOLCHAIN=local` |
The ConfigKit vanity endpoint and direct Go module discovery resolve v1.1.0 to
published commit `191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9`. Its module declares
Go 1.27.0 and retains the YAML dependency and LGPL-3.0-or-later library grant.
The upstream [v1.0 migration notes](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/docs/releases/v1.0.0.md)
and [v1.1 release notes](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/docs/releases/v1.1.0.md)
describe the adopted API and behavior changes. Upgrading from v0.8.0 also crosses
the import, collection, CLI-assignment, and discovery changes documented in the
[v0.9 notes](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/docs/releases/v0.9.0.md)
and [v0.10 notes](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/docs/releases/v0.10.0.md).
A disposable ckconfig copy passed its complete module suite and maintained
simple example tests on Go 1.27.0 after changing the dependency/import path,
minimum, two source-reader fields, and affected test APIs. This establishes that
the existing adapter structure remains suitable. It did not exercise the
proposed report-field renames, additional source evidence, or a migrated
ckprompt; final implementation still needs validation of those selected changes.
### API coverage findings
The follow-up compared the public API of the pinned v1.1.0 module with
ckconfig's provider facets, adapters, and report projections. These are gaps in
current integration coverage, not necessarily APIs newly introduced in v1.1.0.
| Capability | Current ckconfig exposure | Recommendation |
| --- | --- | --- |
| `SourceInfo.EmptyImportSelections` | Omitted from the source report. | Include in adoption; OD-2. |
| `Result.Metadata()`, `Metadata.Lookup`, and `IsExplicit` | No metadata facet in `View`; explanation reports do not expose explicit input presence. | Prioritize optional field-metadata inspection as the next functional expansion; OD-3. |
| `Metadata.HistoricalPaths` and `History` | Operators must already know a path to request an explanation; metadata history records are not exposed. | Pair path discovery and provenance history with metadata inspection; OD-3. |
| `LoadError.Sources()` and `Source(id)` | Acquired sources support diagnostic context, but `config sources` returns before reporting an inventory on acquisition failure. | Consider failed-load source inspection separately; OD-4. |
| Full `Difference` records | The Go comparison provider returns native records; the CLI reduces them to additions, right-side changes, and deletion paths. | Consider an additional detailed comparison report; OD-5. |
| Structured `Diagnostic` records | Native writers present diagnostic content; there is no structured native diagnostic report. | Lower-priority candidate, coordinated with shared reporting; OD-6. |
The upstream [metadata API](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/metadata.go),
[source and diagnostic records](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/records.go),
[load-error API](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/diagnostic.go),
and [comparison API](https://forge.maximumdirect.net/base-layer/configkit/src/commit/191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9/diff.go)
own the returned evidence and its semantics.
The current explanation projection covers every field of v1.1.0's
`Explanation`: effective and historical snapshots, origin and inheritance,
reasons, dependencies, and secret-reference names. Improved list-history
behavior can flow through that existing adapter. Metadata queries add distinct
information: explicitness and null state from `FieldInfo`, observed paths
including removed ones, and `Change` records with previous origins and operation
paths. They need not replace the value-oriented explanation report.
`Result.Source(id)` is a convenient lookup of records already available through
the source inventory. A source-ID filter is a possible usability refinement;
it does not inherently require another command or a required lookup method on
every custom source provider. Discovery, import defaults, collection identities,
and field-level resolution recording fit the application-supplied loader.
Source resource contexts and explicit secrets-directory bootstrap primarily
serve application loading and startup; there is no immediate recommendation to
add standard inspection operations for them. A consumer need could motivate
later presentation of their upstream results.
## Targeted end state
### Dependency and toolchain
The [ckconfig module](../../integrations/ckconfig/go.mod) requires
`go.maximumdirect.net/configkit v1.1.0`, and its production code, tests, and
maintained example use that import identity. Raise its consumer minimum to
Go 1.27.0. Retain its core v0.3.0 requirement unless implementation identifies
a concrete need for a newer core API. Core remains independent of ConfigKit
and retains its own Go minimum.
Use a Go 1.27 toolchain for ckconfig dependency preparation and validation.
Update the [maintainer procedure](../release.md#release-toolchain) to select an
available Go 1.27 release and distinguish the per-module consumer minimums.
The existing host-native, workspace-off launcher can use that toolchain without
a new toolchain-selection mechanism.
### Faithful adaptation and reports
Retain the existing scope, loader, comparison, diagnostic, and reporting
structure. Continue using ConfigKit's public results, native rendering, and
diff APIs rather than reproducing its domain logic locally. New collection,
discovery, and resolution features can be used by application loaders through
the adopted library; they do not inherently require additional CLI flags or
CommandKit-owned loading policy.
Align the existing source report with ConfigKit's current vocabulary:
| Current API or report spelling | Target |
| --- | --- |
| `SourceInfo.DestinationUnresolved` | `SourceInfo.DestinationIsFallback` |
| `SourceInfo.VersionCheck` / `SourceVersionCheck` | `SourceInfo.VersionEvidence` / `SourceVersionEvidence` |
| Text/YAML `destination_unresolved` | `destination_is_fallback` |
| Text/YAML `version_check` | `version_evidence` |
| Native main layer value `base` | Native main layer value `main` |
Update DTOs and both report formats together. Preserve the upstream evidence
and its supported presence, null, location, ordering, and redaction information
through adaptation. Use the upstream terms directly in help and documentation;
link to the owning inspection contract for their semantics. Describe the report
migration plainly in the release note. The target uses the new field names
without duplicate legacy report fields.
Include `EmptyImportSelections` as `empty_import_selections` in both text and
YAML source reports under OD-2. Project each declaration path, location, and
native selector in the returned record order. Emit `empty_import_selections: []`
when there are no records. The [CLI reference](../cli.md#list-configuration-sources)
owns the implemented report contract.
### Deferred inspection expansion
Metadata inspection and path discovery are the next functional work proposed
after adoption (OD-3). Failed-load source inventory, detailed comparison,
and structured diagnostics remain separate later proposals (OD-4 through OD-6).
The [configuration inspection extensions](deferred.md#configuration-inspection-extensions)
record owns their proposed capabilities and design considerations. None is part
of this adoption scope or a new requirement on existing providers.
### Consumer migration and module coordination
Public ckconfig signatures expose ConfigKit types, so switching import identity
requires consumers to migrate the ConfigKit values they pass across those
boundaries. Use the adopted upstream API names in maintained consumers and
fixtures. Known test changes include `Definition.Secret` to `Sensitive` and
`Resolution.Record(..., ChangeNormalization, ...)` to `RecordNormalization`.
Review other removed pre-v1 symbols where actually used.
Migration guidance should link to upstream for source-root `configkit` controls
becoming `config`, `optional` becoming `required` with changed defaults,
identified-collection registrations and operations, and CLI-assignment ordering.
Change affected maintained fixtures when present. Application configuration and
validation choices remain application-owned; CommandKit does not introduce a
compatibility loader or impose new configuration restrictions.
ckprompt's current ConfigKit types come from the old package identity. It cannot
simply select the new ckconfig release while retaining that wiring. Its later
adoption needs a coordinated ConfigKit/PromptKit migration and review of its
own adapters and composed example. Under OD-1, prepare ckconfig independently
first and review ckprompt next. This roadmap does not select an unreviewed
PromptKit release or claim a new matched integration set.
### Independent module compatibility
Once local ckconfig uses the vanity ConfigKit types, the usual three-module
`go.work` selects it for local ckprompt as well. Until ckprompt migrates, its
old-path ConfigKit inspection and comparison types cannot satisfy the new
ckconfig contracts. This is a temporary composition mismatch, even though each
module can be validated independently against its own dependencies.
During that interval, each module's independent validation uses its own declared
dependencies with workspace discovery off. The unchanged ckprompt and composed
example retain published ckconfig v0.2.0 and their existing upstream libraries;
their success does not establish compatibility with migrated local ckconfig.
Development and consumer guidance explain the temporary mismatch. Existing
ignored workspace files remain local; restoring aligned workspace composition
belongs to the later ckprompt migration.
### Release preparation
The target candidate is ckconfig v0.3.0, with full tag
`integrations/ckconfig/v0.3.0`, subject to the existing fresh-version guards.
A pre-v1 minor fits the intentional ConfigKit type-identity, consumer-minimum,
and report-schema changes. Its checked-in note explains these changes and
consumer migration, with the usual immutable licensing footer. Current module
notices describe the selected graph and pinned upstream source.
Adoption includes release preparation, not source publication or hosted
delivery. The existing [release procedure](../release.md) owns those operations
and their separate outcomes when requested; this target adds no release gate.
## Completion evidence
Implementation is complete when ckconfig's actual module metadata and imports
select the new dependency, the selected report changes faithfully expose its
returned evidence, the module suite and maintained consumer pass independently,
and current documentation, notices, and release preparation describe that state.
Completion evidence distinguishes independent module validation from the
temporary workspace composition mismatch. OD-3 through OD-6 remain in deferred
work; their implementation is not a completion condition for adoption. A future
source release is reported separately from implementation completion.
## Selected scope decisions
These dispositions record the user's selected adoption scope.
They record the selected target and do not authorize publication. Deferred
features still need focused API and CLI design.
| Decision | Selected disposition | Reason |
| --- | --- | --- |
| OD-1: Integration sequencing | Implement and prepare ckconfig independently first; review ckprompt next. | Independent releases keep adoption focused. Document the temporary workspace mismatch and lack of a newly aligned release pair. |
| OD-2: Empty-import evidence | Include `empty_import_selections`, with `[]` when empty, in text and YAML source reports. | Exposes existing upstream evidence alongside the other report-schema changes. |
| OD-3: Metadata and path discovery | Defer to the next functional work after adoption. | A coherent optional capability needs its own field/history report and command design. |
| OD-4: Failed-load source inventory | Defer to a separate source-inspection enhancement. | Requires source evidence on error and deliberate report/failure coordination. |
| OD-5: Detailed comparison | Defer; retain the existing reduced report during adoption. | A richer form needs a report-selection and compatibility decision. |
| OD-6: Structured diagnostics | Defer; retain native diagnostic rendering during adoption. | Coordinate machine-readable records with shared reporting and outcome delivery. |
Adoption therefore covers dependency and API migration, source-report
terminology and empty-import evidence, documentation and notices, independent
module validation, and ckconfig release preparation. Core retains its own Go
minimum. Application loaders continue to own discovery and loading policy.
Coordinating both integrations before release remains an alternative if a
matched adoption path becomes necessary. It would expand the scope to include
PromptKit review and ckprompt migration; publication would still proceed in
dependency order. The [deferred feature record](deferred.md#configuration-inspection-extensions)
retains the capability proposals and unresolved design choices for OD-3 through
OD-6, plus the possible source-ID filter.

View File

@@ -1,22 +1,57 @@
# Deferred work
# Future work
Status: retained future work and unselected proposals; no implementation schedule.
Status: pending release and integration work, plus future capability proposals.
No delivery dates are selected here.
## Purpose and authority
Completed initial delivery, performance, ReleaseKit upgrade, and remediation
records have been retired. This document preserves remaining deferred
capabilities and unselected ideas under the
Completed roadmaps, implementation plans, and review findings have been retired.
This document preserves unfinished work and unselected ideas under the
[release policy](../policy/release.md#release-documentation-and-preparation).
The original records remain available in
[Git history](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/6bc37e7ba64ca0920b4d0cc77039a18d4006b24e/docs/roadmap).
The completed ConfigKit adoption, reporting, dependency alignment, release-tooling,
and remediation records remain available at the
[retirement snapshot](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/260256d722f52bc947103d078cb67a129f4cc9f1/docs/roadmap).
[Earlier retired records](https://gitea.maximumdirect.net/eric/CommandKit/src/commit/6bc37e7ba64ca0920b4d0cc77039a18d4006b24e/docs/roadmap)
also remain in Git history. Completed records are not assignments to restart work.
Accepted [ADRs](../adr/) and the [architecture policy](../policy/architecture.md)
govern any future implementation. The [CLI reference](../cli.md), public GoDoc,
and [consumer guide](../consumers/guide.md) own implemented behavior. Retaining
an idea here does not select its syntax, API, priority, or delivery date.
## Deferred capability targets
## Pending release and integration work
### ckconfig release and hosted delivery
The ckconfig v0.3.0 candidate adopts ConfigKit v1.1.0 and Go 1.27.0; its
implementation and audits are complete. Its
[candidate release note](../../integrations/ckconfig/docs/releases/v0.3.0.md)
is prepared, but source publication and hosted delivery remain pending. Complete
those operations when requested through the [release procedure](../release.md),
using the existing fresh-version guards and reporting source publication and
hosted Gitea delivery separately. Publishing this roadmap cleanup does not
publish a module version tag or deliver a release page.
### ckprompt dependency migration and workspace alignment
Review the upstream PromptKit vanity release and select compatible ConfigKit
and PromptKit dependencies before migrating ckprompt's adapters and maintained
composed consumer. The new ckconfig uses vanity ConfigKit types; ckprompt v0.2.0
still uses published ckconfig v0.2.0 and old-path ConfigKit v0.8.0. Their local
sources do not currently compose in the three-module workspace. See the
[consumer guide](../consumers/guide.md#module-selection) for the current module
relationship.
The follow-up includes migrating affected imports and public type boundaries,
selecting real released dependency versions and the required Go minimum,
updating notices and migration guidance, and validating both ckprompt's
independent graph and the aligned composed consumer. Restore three-module
workspace guidance when those dependencies align, updating the ignored workspace
Go version as needed. Until then, preserve local workspace files and validate
with workspace discovery off as described in the [development guide](../development.md).
No unreviewed PromptKit version or new compatibility bridge is selected here.
## Future capability targets
### Comparison extensions
@@ -27,8 +62,8 @@ relevant ADRs.
### Configuration inspection extensions
The [ConfigKit v1.1.0 adoption review](configkit-v1.1.0.md#selected-scope-decisions)
recommends the following work after dependency and source-report adoption.
The [completed ConfigKit v1.1.0 adoption review](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/260256d722f52bc947103d078cb67a129f4cc9f1/docs/roadmap/configkit-v1.1.0.md#selected-scope-decisions)
retained the following work after dependency and source-report adoption.
Metadata inspection is the recommended next functional sprint; the other
features have no selected sequence or delivery date. These proposals do not
establish command grammar, report schemas, or new provider requirements.
@@ -51,7 +86,7 @@ ConfigKit's public APIs own the evidence and its semantics.
failures together; current diagnostic context remains available meanwhile.
- **Detailed comparison (OD-5).** Consider an additional form projecting native
`Difference` paths, both snapshots, presence/redaction fields, and
`HiddenChanged`. The adoption sprint retains the reduced
`HiddenChanged`. The current comparison retains the reduced
additions/changes/deletions report. Decide later whether a detailed form
complements it or replaces it through a deliberate compatibility change.
- **Structured diagnostics (OD-6).** Consider machine-readable native codes,
@@ -83,7 +118,28 @@ A later design should determine how alias suggestions are presented alongside
canonical suggestions while ensuring inserted continuations form valid next
steps. No delivery date or final API is selected here.
## Other deferred proposals
### Reporting extensions
Error-conversion helpers such as `WarnError` remain deferred pending demonstrated
value; applications can already use `report.Warn(err.Error())`. Terminal redraw
and progress-bar presentation were outside the completed reporting scope and
remain unselected. Whole-run buffering to make failed runs show only errors
was also outside that scope. Any later extension needs a deliberate API and
presentation decision rather than assuming those behaviors are already provided.
### Release workflow extensions
Historical hosted-page backfilling and automated CI delivery remain separate
future choices. The current local page-only delivery workflow is implemented;
these extensions are not prerequisites for using it. The release procedure owns
delivery, immutable content, and recovery behavior.
Additional validation options, including race testing, full vet, compilation
matrices, size limits, and more forbidden-file paths, can be considered for
demonstrated needs. None is selected by retaining it here, and the existing
candidate checks do not become general prohibitions on development workspaces.
## Other future proposals
These capabilities remain outside the delivered scope and require a concrete
operator or consumer need and an appropriate ownership/behavior decision:

View File

@@ -1,332 +0,0 @@
# ConfigKit v1.1.0 adoption implementation plan
Status: Stages 1–6 complete.
This plan is for a Luna coding agent implementing the selected
[ConfigKit adoption target](configkit-v1.1.0.md). Follow Stage 1 through Stage 6
in order as one end-to-end sprint. Intermediate editing states need not compile
or form independently usable releases. Reach and validate the complete selected
end state before claiming completion; do not introduce compatibility machinery
merely to make an intermediate stage pass.
This planning pass changes documentation only. Execution prepares a ckconfig
release candidate; source publication, hosted delivery, ckprompt migration, and
the deferred inspection extensions are outside this sprint. Do not push or
create release tags as part of these stages. The user's later publication
instructions and the [release procedure](../release.md) govern those operations.
Start with [development guidance](../development.md) and follow the
[agent](../policy/agents.md), [architecture](../policy/architecture.md),
[testing](../policy/testing.md), [documentation](../policy/documentation.md),
[release](../policy/release.md), and [OSS](../policy/oss.md) policies for the work
they govern. Use the repository's codebase-memory tools for code discovery.
The completed [vanity migration plan and publication evidence](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/16584084b1b8a87dac959334ede8a47a8cfb8c0b/docs/roadmap/implementation.md)
remain in Git history; do not restart that work or the completed ReleaseKit
adoption/remediation.
## Baseline and fixed choices
Source/review baseline: `16584084b1b8a87dac959334ede8a47a8cfb8c0b`.
The planning commit changes roadmap documents, not module code.
| Item | Baseline | Selected end state |
| --- | --- | --- |
| ckconfig module | `go.maximumdirect.net/ckconfig`, published v0.2.0 | Same module identity; prepare v0.3.0 |
| ConfigKit | `gitea.maximumdirect.net/eric/ConfigKit v0.8.0` | `go.maximumdirect.net/configkit v1.1.0` |
| ckconfig Go minimum | Go 1.26.0 | Go 1.27.0 |
| Core dependency | `go.maximumdirect.net/commandkit v0.3.0` | Retain v0.3.0 |
| Maintainer toolchain | Go 1.26.8 | Go 1.27.0 for preparation, validation, and ReleaseKit execution |
| Other module minimums | Core and ckprompt: Go 1.26.0 | Retain their independent minimums |
ConfigKit v1.1.0 is published at
`191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9`. Its module requires Go 1.27.0;
the prior disposable migration passed ckconfig and its example tests on that
toolchain. The selected report changes have not yet been implemented or tested.
The feature roadmap's selected OD-1/OD-2 choices apply: migrate ckconfig first
and include empty-import-selection evidence. OD-3 through OD-6 and source-ID
filtering remain in [deferred work](deferred.md#configuration-inspection-extensions).
Retain the existing configuration operations, provider facets, loading and
comparison coordination, diagnostics, outcomes, and reduced diff report.
ConfigKit owns domain semantics; this sprint adapts its public capabilities.
Retain root ReleaseKit v0.9.0, the release launchers and profiles, physical
module directories, and integration tag prefixes. Core and ckprompt code and
module selections are outside this adoption. Published old releases and notes
remain immutable. A new adapter layer, legacy import bridge, duplicate legacy
report fields, or local reimplementation of ConfigKit semantics is unnecessary.
Preserve ignored `go.work`, `go.work.sum`, and `.releasekit/gitea-token` without
reading credential contents. Use `GOWORK=off GOTOOLCHAIN=local` throughout module
preparation, validation, and example runs. Local ckprompt will not compose with
migrated local ckconfig until its later ConfigKit/PromptKit migration; its
independent suite continues to use published ckconfig v0.2.0. Do not edit
ckprompt or introduce replacements to conceal this temporary mismatch.
## Stage 1: Migrate the dependency, imports, and removed Go APIs
**Completed.** Ckconfig now selects ConfigKit v1.1.0 and Go 1.27.0, with
authentic downloaded sums and migrated imports and native API accesses. An
offline compile of its packages and example passed with Go 1.27.0.
1. Inspect the worktree and preserve unrelated work. Select Go 1.27.0 on
`PATH` and confirm it with `GOWORK=off GOTOOLCHAIN=local go version`.
This version was available during planning. Reuse an existing installation,
or install/download `golang.org/dl/go1.27.0` using the established SDK setup
pattern in the release procedure. Do not change the ignored workspace to
select the toolchain.
2. In `integrations/ckconfig/go.mod`, set `go 1.27.0`, remove the old ConfigKit
requirement, and require `go.maximumdirect.net/configkit v1.1.0`. Keep the
core requirement and YAML selection unless ordinary resolution establishes
a required change. Do not tidy root/ckprompt merely to raise their minima.
3. Migrate ckconfig-owned Go imports, including tests and the simple example,
to the new ConfigKit identity. Preserve the package alias `configkit`.
Inspect literals and documentation references separately from imports;
old repository URLs in published history are not import-path defects.
4. Adapt the native source-field accesses in `sources.go` to
`DestinationIsFallback` and `VersionEvidence`. Use
`configkit.SourceVersionEvidence` in affected test fixtures. The public
`SourceReader` contract still returns `[]configkit.SourceInfo` through the
new import identity; it needs no additional methods.
5. Update removed APIs actually used by the tests: `Definition.Secret` becomes
`Sensitive` in `comparison_adapter_test.go`, `diff_execution_test.go`, and
`diff_report_render_test.go`; the normalization call in
`TestFromConfigKitCompareIgnoresDefaultAndResolutionHistoryForEqualValues`
becomes `RecordNormalization(path, reason)`. Preserve the existing
application mutation and assertions. Search for any other used removed
symbols and use the pinned upstream migration mapping, not aliases or
replacement domain logic.
6. After import migration, prepare dependencies and run `go mod tidy` from
ckconfig with workspace discovery off and the selected local toolchain.
Use normal dependency retrieval during preparation; inspect the authentic
resulting `go.mod`/`go.sum` changes. Do not fabricate sums or use a sibling
checkout as a dependency. Format changed Go files.
**Stage result:** ckconfig selects the released vanity dependency and current
Go APIs. Report spelling and new evidence are completed by the next stages;
there is no requirement to keep this intermediate state independently usable.
## Stage 2: Align source-report vocabulary and add empty-import evidence
**Completed.** Text and YAML projections use the new vocabulary and include
ordered empty-import selections, including empty lists and nil version evidence.
The packages compile offline with Go 1.27.0; focused behavioral tests follow.
Work in `integrations/ckconfig/sources.go`. Keep its existing report preparation
and delivery path; no new provider capability or operation is needed.
1. Rename DTO fields/types and both text/YAML keys to match the upstream terms:
`DestinationIsFallback` / `destination_is_fallback`, and
`VersionEvidence` / `version_evidence`. Rename `sourceVersionCheckDTO` to
`sourceVersionEvidenceDTO`. Continue projecting all existing version,
location, presence, null, and redaction fields as returned. Allow the native
main layer value to appear as `main`; do not translate it back to `base`.
2. Add `EmptyImportSelections` to each source DTO, serialized as
`empty_import_selections` after `version_evidence`. Use an ordinary slice of
a small DTO with fields in this order: `path` string, `location` using the
existing `locationDTO`, and `selector` string. No `omitempty` is needed:
initialize the slice so sources without records serialize as `[]`.
3. For every native record, project `Path.String()`, its supplied location's
source ID/line/column, and `string(Selector)`. Preserve source and record
order, source-local coordinates, empty strings, and zero coordinates.
Forward selector strings without validating or remapping them. Apply no
additional import discovery, filesystem access, or domain interpretation.
4. Extend the text writer with the same field names and field order. For an
empty slice, print `empty_import_selections: []`. For nonempty slices, print
one indented record per entry with `path`, `location`, and `selector`.
Follow the existing text style: quote path/source-ID/selector strings with
`strconv.Quote` and render a location as
`{source_id: "...", line: N, column: N}`. YAML uses the equivalent mapping.
5. The current text writer uses `continue` after printing a null version check.
Restructure that conditional so a null `version_evidence` does not skip the
new empty-import field. Every reported source includes both fields, with
`version_evidence: null` where the upstream pointer is nil.
**Stage result:** both report forms use the current vocabulary and expose
empty-import evidence, including when no version evidence exists. Existing
source inventory, explanation, diagnostic, and outcome semantics are retained.
## Stage 3: Update focused report and adapter coverage
**Completed.** Focused source tests cover the new text and YAML keys, native
main layer, ordered selections, empty lists, and nil/redacted evidence. The
offline `TestSources` suite passed with Go 1.27.0.
1. Extend `TestSourcesProjectNativeEvidenceInTextAndYAML` in `sources_test.go`
rather than building another report-test framework. Update fixtures and
assertions to the new native fields and report keys. Check that legacy
`destination_unresolved` and `version_check` keys are absent from emitted
source records; existing published history need not be scanned away.
2. Add controlled empty-import records to the source-reader fixture. Use
multiple records with distinguishable paths, selectors, and locations so
the test detects loss, reordering, or incorrect projection. Include a
source with nil version evidence and nonempty import records to cover the
text-writer conditional, and a source with no import records to verify `[]`
in both forms. Preserve the existing null/redacted version-evidence cases,
source order, zero-value coverage, and inventory-read count assertions.
3. Use the native `LayerMain` value in an appropriate fixture and verify that
returned source/explanation layer names are presented faithfully. Reuse
existing assertions where they already cover native origin projection;
do not add a separate test of ConfigKit's layer or provenance rules.
4. Keep existing tests of loader request forwarding, comparison coordination,
diagnostics, cancellation, error identity, output failures, and safe native
snapshot adaptation. Adjust removed APIs or intentionally changed report
expectations only. Do not rewrite behavior assertions merely to make an
unexpected regression pass.
5. Run the focused source tests with prepared dependencies, from ckconfig:
```sh
GOWORK=off GOTOOLCHAIN=local GOPROXY=off GONOPROXY=none GOSUMDB=off \
go test -mod=readonly -count=1 -run '^TestSources' .
```
**Stage result:** durable coverage establishes CommandKit's new report
projection. The fake native inventory supplies evidence; tests do not establish
whether ConfigKit discovers imports, checks versions, matches records, or
classifies sensitive data correctly.
## Stage 4: Update current guidance and prepare the release note
**Completed.** Current guidance describes the new source report, consumer and
toolchain migration, and temporary ckprompt mismatch. The v0.3.0 candidate
note is prepared; published notes remain unchanged.
1. Update ckconfig README, GoDoc, and `examples/simple/README.md` for the new
dependency identity and Go minimum, and the planned ckconfig v0.3.0 with
core v0.3.0. Keep package/provider usage consistent with actual code.
Review maintained fixtures for source controls or removed APIs; migrate
genuine old controls where present, without changing application fields
merely because they have similar names.
2. Update `docs/cli.md`'s configuration-source section with the renamed keys,
main layer value, and `empty_import_selections` schema from Stage 2,
including the empty-list shape. Update origin-layer examples where they
describe adopted ConfigKit results. Retain existing command grammar,
formats, status behavior, and reduced comparison report. Use upstream terms
directly and link to the pinned upstream contracts for their semantics.
3. Update `docs/consumers/guide.md` and README's module-selection summary to
distinguish two cases: the newly prepared ckconfig v0.3.0 with ConfigKit
v1.1.0, and the existing matched ckconfig/ckprompt v0.2.0 pair. Do not imply
that ckprompt v0.2.0 composes with new ckconfig. Qualify upstream links where
ckconfig and unchanged ckprompt now use different ConfigKit versions.
4. Document the consumer migration: ConfigKit import/type identity, Go 1.27.0,
changed report keys, and added source evidence. Link to upstream migration
guidance for source-root `configkit` controls becoming `config`, import
requiredness/defaults, collection registrations/operations, removed public
APIs, and CLI-assignment order. Keep the detailed configuration reference
upstream rather than creating a parallel CommandKit contract.
5. In `docs/release.md`, select Go 1.27.0 as the maintainer toolchain and adapt
its existing install/download/PATH example. In that procedure and the
consumer guide, distinguish ckconfig's Go 1.27.0 minimum from core and
unchanged ckprompt's Go 1.26.0 minima. Keep ReleaseKit v0.9.0 and its
existing launcher, profiles, guards, publication, delivery, and recovery
workflow. This does not require code or profile changes.
6. Replace the development guide's unqualified three-module workspace advice
with interim guidance: preserve ignored workspace files and use
`GOWORK=off GOTOOLCHAIN=local` for preparation, suites, and example runs.
Explain that ckprompt's independent suite and composed example use its
published ckconfig v0.2.0 dependency; full local workspace composition is
pending the later ckprompt migration. Restore aligned workspace guidance
during that later work, including the workspace Go version as needed.
7. Create `integrations/ckconfig/docs/releases/v0.3.0.md`, subject to the
existing fresh-version selection, explaining the complete v0.2.0-to-candidate
changes and migration actions. End with `COPYRIGHT | LICENSE | ATTRIBUTION`
links to the corresponding module-local paths under
`https://forge.maximumdirect.net/base-layer/commandkit/src/tag/integrations/ckconfig/v0.3.0/integrations/ckconfig/`.
Future tag links need not resolve before publication. Preserve old notes.
**Stage result:** current owners describe implemented adoption and the interim
module relationship; a checked-in candidate release note is ready. No deferred
feature, release publication, or external application migration is claimed.
## Stage 5: Align module attribution with the actual selected graph
**Completed.** The workspace-off module graph selects ConfigKit v1.1.0 and
vanity ReleaseKit v0.9.0; old-path ReleaseKit is no longer selected. Attribution
now reflects that graph and the pinned ConfigKit copyright and license source.
1. From ckconfig, inspect `GOWORK=off GOTOOLCHAIN=local go list -m -json all`
after tidy. Compare it with the existing module attribution. ConfigKit's
own tool requirement does not select CommandKit's release tool: retain the
root-selected ReleaseKit v0.9.0 and inspect the effective transitive graph.
2. Update `integrations/ckconfig/docs/oss/attribution.md`'s ConfigKit entry to
`go.maximumdirect.net/configkit v1.1.0` and pinned commit
`191015e289d7f2b91abb9fdd9c4f2f5cf75b4de9` on the new forge. Verify its
actual copyright, grant, license, and applicable notices. Reuse existing
complete standard license texts when they match the selected terms.
3. Remove/update entries for dependencies no longer selected or whose selected
versions changed; the baseline includes old-path ReleaseKit v0.4.1 and its
dependencies as well as vanity ReleaseKit v0.9.0. Follow the actual graph,
not a global old-host replacement or an assumption that every tool module
disappears. Keep the House Format and exact-revision notice/source links.
4. Core and ckprompt dependency selections remain unchanged. Preserve their
corresponding notices and published release records. Update shared material
only if its actual scope or selection changes; adoption alone does not
require revising unrelated grants or copying upstream policy documents.
**Stage result:** ckconfig's notices correspond to its real selected modules,
with authentic source/license references and no unrelated attribution rewrite.
## Stage 6: Validate the final state and finish the sprint
**Completed.** With Go 1.27.0 and workspace discovery off, the complete core,
ckconfig, and ckprompt suites passed offline with read-only metadata. Ckconfig's
suite was repeated after adding an explicit native main-layer explanation
assertion. The simple example's text and YAML source reports and fixture diff
ran successfully; the focused source fixture covers nonempty import evidence.
Core and ckprompt were tested against their own declared dependency graphs,
not against local migrated ckconfig. Full local workspace composition remains
pending the later ckprompt migration. The v0.3.0 note is a candidate; no source
release or hosted delivery was performed.
1. Format changed Go files; inspect final diffs and module metadata. Confirm
ckconfig-owned Go imports use the vanity ConfigKit identity, the old
requirement is gone, its Go minimum is 1.27.0, core stays at v0.3.0, and no
local replacements/workspaces enter tracked module inputs. Confirm root
and ckprompt still have their existing module selections and Go minima.
2. Prepare any missing dependency caches for all three modules using normal
retrieval with workspace discovery off. On Go 1.27.0, run each independent
complete suite with read-only offline dependencies:
```sh
GOWORK=off GOTOOLCHAIN=local GOPROXY=off GONOPROXY=none GOSUMDB=off \
go test -mod=readonly -count=1 ./...
(cd integrations/ckconfig && \
GOWORK=off GOTOOLCHAIN=local GOPROXY=off GONOPROXY=none GOSUMDB=off \
go test -mod=readonly -count=1 ./...)
(cd integrations/ckprompt && \
GOWORK=off GOTOOLCHAIN=local GOPROXY=off GONOPROXY=none GOSUMDB=off \
go test -mod=readonly -count=1 ./...)
```
This includes ckconfig's maintained example tests. Record the unchanged
modules' results as validation of their own published dependency graphs,
not composition with local migrated ckconfig. Do not add a workspace shim
or change ckprompt to obtain a green three-module workspace.
3. Exercise the simple example's `config sources` text/YAML reports and its
existing fixture comparison with workspace discovery off. Use a controlled
native source fixture from Stage 3 for nonempty import evidence; a new
application import system or live service is unnecessary. Confirm only the
selected presentation and wiring behavior, not upstream domain correctness.
4. Check affected current documentation, local links/fragments, whitespace,
candidate note paths, attribution entries, and pinned-source provenance.
Ensure future behavior is not described as implemented, and distinguish
prepared v0.3.0 from a published release. Run `git diff --check` and the
staged counterpart, including intended new files in the review.
5. Record stage completion, final validation evidence, and the temporary
workspace mismatch in this plan. Mark the feature roadmap's implementation
status accordingly while leaving publication and external adoption pending.
Keep deferred capabilities in their owning record without starting them.
6. Review the worktree and staged changes, commit completed units under the
agent policy with concise messages, and finish with a clean worktree apart
from explicitly preserved unrelated work. Summarize the module/dependency
changes, report migration, tests, commits, and remaining ckprompt follow-up.
Do not publish; the normal selected-module release workflow remains the
owner of future candidate validation, guards, and publication outcomes.
**Stage result:** the selected ckconfig adoption is implemented, documented,
validated, committed, and prepared for a later release request. Intermediate
compile failures and deferred work are not substitutes for this final result.
No required scope questions remain open. Routine implementation choices follow
the selected upstream contracts and the report shape fixed above; deferred
extensions do not need to be designed to complete this sprint.

View File

@@ -1,207 +0,0 @@
# Integration dependency alignment and vanity-path migration
Status: the vanity migration and matched source releases are complete. The
[completed implementation record](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/16584084b1b8a87dac959334ede8a47a8cfb8c0b/docs/roadmap/implementation.md) records candidate commits, validation,
and publication outcomes; this roadmap retains the target and review history.
## Purpose and authority
Align the independently released integrations with the core APIs their current
source already uses, and migrate all three CommandKit modules to the supplied
vanity paths. Publish compatible module versions in dependency order so each
integration builds and validates against its declared, downloaded dependencies
with workspace discovery off.
The [release procedure](../release.md) owns publication commands and release
order. The [release policy](../policy/release.md),
[agent policy](../policy/agents.md), [testing policy](../policy/testing.md),
[documentation policy](../policy/documentation.md), and
[OSS policy](../policy/oss.md) govern the work. This roadmap selects the target
and release-preparation scope; it does not change those policies.
## Baseline and investigation evidence
Reviewed baseline: `289a171e97699ad9aa57ee2cc40bdad1c0b680a9`.
Vanity-path review baseline: `d7cbaed3ed18bc2909c8144f308e2dfd58ade7ec`.
Module declarations and library source were unchanged between these revisions.
- Both integration module declarations require core v0.2.0. The
[ckprompt declaration](../../integrations/ckprompt/go.mod) also requires
ckconfig v0.1.0; the [ckconfig declaration](../../integrations/ckconfig/go.mod)
has no dependency on ckprompt.
- Current integration source uses newer core APIs including `CompletionHint`,
`Explanation`, and `Reporter`. Both independent Go 1.26.8 suites fail to
compile with `GOWORK=off` and their existing read-only module graphs.
- Updating core alone is insufficient for ckprompt: published ckconfig v0.1.0
still uses removed core APIs, including `Invocation.Streams`,
`Outcome.Diagnostics`, and `SeverityError`, plus the old `Failure(error)`
call shape.
- The authoritative `origin` tag listing contains core v0.1.0 and v0.2.0, and
the two integration v0.1.0 tags. The proposed next versions below were unused
at review. Remote `main` was `230187296fae5253cb1773f16c2f1d1a39499ed0`;
the reviewed local changes still need normal publication to the authoritative
branch as part of candidate preparation.
- A checked-in [core v0.3.0 draft](../releases/v0.3.0.md) already describes the
reporting and alias migration. It is an unpublished draft, not proof that
the module version can be downloaded.
The investigation used disposable alternate module files, Go 1.26.8,
`GOWORK=off`, and read-only offline tests. Only temporary dependency selections
changed; the real module declarations, sums, and source remained unchanged.
| Current source under test | Temporary dependency selection | Result |
| --- | --- | --- |
| ckconfig | Current core source | Module suite and maintained example passed. |
| ckprompt | Current core source, published ckconfig v0.1.0 | Compilation failed in the old ckconfig implementation. |
| ckprompt | Current core and current ckconfig source | Module suite and maintained example passed. |
The successful cases retained ConfigKit v0.8.0 and PromptKit v0.16.0. These
results isolate the CommandKit dependency alignment needed here; they do not
prove that future published versions or their downloaded sums already exist.
### Verified vanity discovery
The core endpoint at `https://go.maximumdirect.net/commandkit?go-get=1` serves
the mapping to `https://forge.maximumdirect.net/base-layer/commandkit`. Go direct
discovery finds the existing core v0.1.0 and v0.2.0 tags. Fetching v0.2.0 through
the vanity path reaches its source but reports a module-path mismatch because
that release declares `gitea.maximumdirect.net/eric/CommandKit`.
Both integration endpoints now return HTTP 200 with these `go-import` contents:
| Vanity path | VCS | Repository URL | Subdirectory field |
| --- | --- | --- | --- |
| `go.maximumdirect.net/ckconfig` | `git` | `https://forge.maximumdirect.net/base-layer/commandkit` | `integrations/ckconfig` |
| `go.maximumdirect.net/ckprompt` | `git` | `https://forge.maximumdirect.net/base-layer/commandkit` | `integrations/ckprompt` |
Go 1.26.8 direct discovery finds each integration's v0.1.0 tag. Disposable
consumer fetches reach the correct nested `go.mod` and report its old declared
module path, confirming subdirectory routing. Discovery is working; the
existing tags have not migrated. The optional fourth metadata field is
supported from Go 1.25, within the existing Go 1.26.0 consumer minimum; see the
[Go discovery reference](https://go.dev/ref/mod#vcs-find).
## Targeted end state
### Compatible independently released modules
Use the following proposed next versions, subject to the existing fresh-version
publication guards:
| Module path | Physical module directory | Next release | Required CommandKit dependencies | Full tag |
| --- | --- | --- | --- | --- |
| `go.maximumdirect.net/commandkit` | `.` | v0.3.0 | None | `v0.3.0` |
| `go.maximumdirect.net/ckconfig` | `integrations/ckconfig` | v0.2.0 | `go.maximumdirect.net/commandkit v0.3.0` | `integrations/ckconfig/v0.2.0` |
| `go.maximumdirect.net/ckprompt` | `integrations/ckprompt` | v0.2.0 | `go.maximumdirect.net/commandkit v0.3.0` and `go.maximumdirect.net/ckconfig v0.2.0` | `integrations/ckprompt/v0.2.0` |
Minor integration releases fit the existing pre-1.0 compatibility policy: the
reporting migration changes consumer-facing behavior and adopts a core version
with incompatible API changes. The vanity migration also deliberately changes
Go module and import identities. Migration notes explain the new imports and
dependency set, provider results, invocation reporting, and application changes.
The three modules retain independent versions and source tags. Their candidates
may identify successive commits: core publication need not wait for integration
metadata that depends on that newly published core. Each selected module's
final release suite runs against its actual declared dependencies.
### Module and import identity
Change each module declaration and all CommandKit-owned imports to the matching
new path. Core self-imports, including its `internal` packages, use
`go.maximumdirect.net/commandkit/...`; integration self-imports use
`go.maximumdirect.net/ckconfig/...` or `go.maximumdirect.net/ckprompt/...`.
Cross-module imports and requirements use the complete new paths in the table,
including ckprompt's ckconfig dependency. Core component-version discovery uses
`go.maximumdirect.net/commandkit` as its build-info identity, retaining the
existing `version --scope commandkit` behavior with the new module path.
Update production files, tests, maintained examples, and any import-path-based
test expectations together. A declared vanity module that still imports the
old core path would select a different Go package identity and can violate
`internal` import rules. The matched new releases compose through one core
identity rather than mixing old-path and new-path CommandKit types.
Keep physical directories, repository URLs, and Git tag prefixes as shown. The
Go subdirectory metadata makes the short integration vanity paths resolve to
the existing layout. The existing release profiles select physical directories;
the migration does not rename those directories or flatten integration tags.
Published old-path versions remain immutable. This scope selects the new paths
for new releases and updates consumer guidance; it makes no new promise of
type interoperability between the two identities and adds no old-path forwarding
module or compatibility layer.
### Retained architecture and scope
Preserve exported declaration shapes and runtime behavior while deliberately
changing module/import identities. Retain the integration adapters, Go 1.26.0
consumer minimum, Go 1.26.8 maintainer toolchain, and root-selected ReleaseKit
v0.9.0. The earlier successful diagnostic suites establish the dependency fix
before renaming; the migrated source still needs its own validation. This work
selects no API rollback or shim for old reporting APIs.
Retain ConfigKit v0.8.0, PromptKit v0.16.0, and other dependency selections
unless ordinary tidy changes the resolved graph as a consequence of the
CommandKit updates. Inspect and explain the resulting graph rather than
bundling unrelated upgrades. In particular, transitive tools with different
module paths may coexist; attribution must reflect the actual selection.
Preserve the existing launchers, three release profiles, generic validators,
credential handling, and separate source-publication and hosted-delivery
outcomes. Development workspaces remain useful and ignored. Diagnostic local
replacements and the earlier temporary Go shim remain development evidence;
release validation uses normal downloaded dependencies without that redirection.
## Release-preparation deliverables
The core v0.3.0 draft describes the complete candidate, including completion,
reporting, aliases, tool maintenance, and the vanity-path migration. New
integration notes at `integrations/ckconfig/docs/releases/v0.2.0.md` and
`integrations/ckprompt/docs/releases/v0.2.0.md` explain the matched dependency set
and consumer migration. Their source and licensing links follow the current
release procedure; published notes and tags remain immutable.
Current installation examples, README module selections, consumer/developer
guidance, affected GoDoc, and maintained examples use the implemented new
identities. Import-path-aware tests retain their existing boundary assertions.
Source-hosting and license URLs remain repository links, distinct from Go
import paths. Module-local attribution describes each actual resolved graph.
The final record identifies module paths, versions, candidate revisions,
validation evidence, and independent source-publication outcomes. Current
readiness guidance reflects the aligned state; historical failures remain in
completed investigation records. The implementation plan owns the ordered
migration, dependency preparation, release, and evidence-recording steps.
## Validation and completion boundaries
Review current CommandKit-owned source and dependency selections for stale
self-imports or old direct dependency keys. Historical notes, source URLs, and
unrelated upstream module identities are outside that mechanical rename.
Confirm release identity still pairs each new module path with its existing
physical directory and tag prefix. The verified redirects are preparation
evidence; publishing the renamed modules is needed for real consumers to use
them through those paths.
Use the selected module's ordinary deterministic offline suite with
`GOWORK=off` and read-only metadata after dependency preparation. Its maintained
consumer compiles and runs with that suite. ReleaseKit's publication command
owns the final candidate suite and selected file checks; validation-only checks
remain optional diagnostics. Reuse successful evidence for unchanged candidates
instead of adding a required check-then-publish repetition.
Retrieving a newly published dependency is preparation for the next module's
release. It does not turn proxy availability or consumer installation into a
completion gate for the preceding source release. Continue to resolve failed
or ambiguous publication using the existing policy and upstream recovery flow.
Hosted Gitea delivery keeps its separate outcome; this roadmap adds no CI
polling, live provider calls, or additional publication gate.
Alignment is complete when the matched source versions are published under all
three vanity module paths, owned imports and examples use the new identities,
each integration declares the compatible vanity-path versions with authentic
downloaded sums, each final selected-module suite passes without workspace or
replacement redirection, and current documentation and attribution agree with
that graph. Temporary local-source success alone does not establish that end
state.

View File

@@ -1,198 +0,0 @@
# Release tooling simplification
Status: the selected release-tooling design has been implemented. All four
generic file checks, registered-tool execution, page-only hosted delivery, and
protection of the supplied token path are selected.
The [completed implementation record](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/87eb84ccb462bf54e6250c8b287ac217efe90883/docs/roadmap/implementation.md)
contains stage order, tasks, validation, and completion evidence. Current release
behavior stays with its canonical owners.
## Purpose and authority
Reduce CommandKit's release infrastructure to declarative module profiles,
a small launcher, and focused tests of CommandKit-owned wiring. Delegate
reusable candidate validation and publication mechanics to ReleaseKit while
retaining CommandKit's domain-specific checks in their existing owners. Add
hosted Gitea release pages as a separate local delivery operation after source
publication.
The [agent policy](../policy/agents.md),
[release policy](../policy/release.md#candidate-validation), and
[testing policy](../policy/testing.md#upstream-integration-testing-ownership)
govern implementation. The [release procedure](../release.md) continues to own
the current workflow.
## Targeted end state
### Ownership
CommandKit retains module selectors, product identities, profile locations,
the authoritative endpoint selection, release order, release-note meaning,
compatibility decisions, and attribution accuracy. Its existing public API,
CLI, integration-adaptation, and consumer-example tests remain local and run
through the owning module's suite. No parallel release-only domain suite is
needed unless a concrete CommandKit requirement is otherwise unprotected.
ReleaseKit supplies the standard offline candidate suite, selected generic file
checks, candidate/note and publication guards, tag creation and publication,
hosted delivery, outcome reporting, and recovery. CommandKit's launcher delegates
those operations. Existing source-release completion and immutability rules
continue to apply. Hosted pages are the added delivery channel; binary
deliverables and platform support commitments remain unchanged.
### Profiles
Select these existing ReleaseKit options in each profile:
| Selection | Intended protection |
| --- | --- |
| `validation.go_format: true` | Detect unformatted committed Go source. |
| `validation.markdown: {}` | Detect broken local Markdown destinations and supported heading fragments. |
| `validation.module_distribution.workspace_files: true` | Detect committed development workspace files in selected inputs. |
| `validation.module_distribution.local_replacements: true` | Detect filesystem replacements in selected committed module declarations. |
Also select
`validation.forbidden_files: [".releasekit/gitea-token"]` in all three profiles.
This delegates protection of the supplied credential path to ReleaseKit's
existing check. The path is repository-relative and the check applies across
the captured repository even for an integration release. It does not read token
contents, scan other paths, or inspect Git history.
Retain the explicit shared-file lists and use ReleaseKit's existing input scope.
The root profile does not run every nested module's suite. File checks inspect
committed inputs and do not replace formatting worktree changes or reviewing
documentation meaning, attribution, and external links. Markdown keeps its
default treatment of outside-checkout destinations; no stricter boundary mode
is proposed.
The pinned [ReleaseKit configuration reference](https://forge.maximumdirect.net/base-layer/releasekit/src/commit/23203f235cc414c4ffe0fd3941030eb0c9f9943f/docs/config.md)
owns exact scope and behavior, including nested-module boundaries and vendored
inputs. Do not expand an option's scope into a general repository prohibition.
Race testing, full vet, compilation matrices, size limits, and additional
forbidden-file paths are not selected by this roadmap. They can be considered
separately for demonstrated needs.
### Launcher and recovery
Use the root module's registered `go tool releasekit` through `exec`, keeping
the current selector interface and publication argument mapping.
Retain host normalization, workspace-off execution, the local toolchain, and
read-only module metadata. Preserve other effective Go settings while arranging
read-only selection.
The implementation plan specifies the wrapper interface and invocation details.
Go's cache supplies executable storage; the launcher creates no persistent
binary directory. `/.releasekit/` remains ignored for the credential file, with
an explanatory comment. Recovery instructions use the registered tool or the
entry point printed for the saved attempt, preserving ReleaseKit's distinction
between publication outcome and record-storage errors.
Share module selection and host-safe tool invocation between publication,
delivery, and recovery where applicable. Delivery selects the same module
profile as publication; recovery uses its recorded attempt rather than inferring
identity from a current profile. Keep the existing publication interface and
provide an explicit delivery invocation without automatically chaining the two
operations. Keep one selector mapping and shared invocation plumbing.
ReleaseKit v0.9.0 shipped host-safe consumer launcher examples and regression
coverage. The earlier simplification used v0.8.0, which already supported
registered tools. Keep host normalization locally because the reviewed
`go tool` invocation inherited foreign `GOOS`/`GOARCH` values and failed to run.
Do not switch tool-version selection from the root module to each integration
merely to match the upstream multi-module example.
### Hosted Gitea release pages
Configure this delivery selection in each module profile:
| Field | Selected value |
| --- | --- |
| `delivery.forge` | `gitea` |
| `delivery.base_url` | `https://forge.maximumdirect.net` |
| `delivery.owner` | `base-layer` |
| `delivery.repository` | `commandkit` |
| `delivery.token_file` | `.releasekit/gitea-token` |
Preserve the profiles' distinct product identities and physical module
directories. All pages belong to the same Gitea repository and are keyed by
the complete tag: `vX.Y.Z` for core, `integrations/ckconfig/vX.Y.Z` for ckconfig,
and `integrations/ckprompt/vX.Y.Z` for ckprompt. Pass the selected profile and
semantic version to `deliver`; ReleaseKit derives the complete module tag.
With no build targets, delivery is page-only. ReleaseKit uses the complete
checked-in module release note, including its licensing footer, as the hosted
body. No binary assets, receipts, or checksum manifests are added. Use current
forge URLs and the full module tag for future notes' immutable source and
licensing links. Existing published tags and notes remain unchanged.
After source publication succeeds, explicitly invoke delivery from a clean
checkout whose `HEAD` identifies that published tag's commit. For later retries,
use the original tagged source and its recorded module files. Report the source
publication outcome and hosted-delivery outcome separately, including the hosted
URL on success. A delivery failure leaves the source release complete; repair
delivery without republishing or moving its tag.
Delivery performs its own read-only credential preflight, without rerunning
candidate validation or publishing a tag. Hosted-service availability and the
Gitea token do not become source-publication gates. A nonblank
`RELEASEKIT_GITEA_TOKEN` environment value takes precedence over the configured
file. The file path resolves from the repository root for all three modules.
ReleaseKit accepts matching hosted content and reports conflicting content
without overwriting it. Resolve uncertain hosted writes before retrying the
same immutable version. The pinned
[delivery guide](https://forge.maximumdirect.net/base-layer/releasekit/src/commit/23203f235cc414c4ffe0fd3941030eb0c9f9943f/docs/consumers/delivery.md)
owns exact delivery behavior and recovery. Start with tags containing the new
delivery configuration; historical page backfilling remains a separate decision.
This local workflow needs no CI delivery job or additional secret distribution.
### Focused local tests
Replace the release-test scaffolding with a small controlled Go-command boundary
that invokes the actual launcher. Cover:
- Selector-to-profile/module routing for core, ckconfig, and ckprompt, including
consistency with the checked-in profiles' intended module and product.
- Invalid argument handling before tool invocation.
- Invocation from another directory and paths containing spaces.
- Host normalization and workspace/toolchain/read-only settings.
- Publication and delivery argument, output-stream, and child exit-status
forwarding, with the selected module profile reaching the delivery operation.
- Profile wiring for the shared Gitea destination, token path, and its selected
committed-path protection; assertions do not read the credential file.
Exercise module routing separately from exit statuses rather than multiplying
every combination. Assert behavior and meaningful selections, not helper
implementation details or exact usage prose.
The resulting local suite omits the disposable publication exercise, generic
tag inspection, assertion-helper self-tests, and custom build-call log/parser.
Upstream publication and delivery coverage owns those generic behaviors. No
replacement publication engine or exported test-helper API is needed. There is
no line-count target for the replacement tests.
## Selected choices and rationale
All four generic file checks are selected as release-time checks, in addition
to the agreed forbidden token path. Formatting and local-link validation automate
existing contributor review work. Workspace-file and filesystem-replacement
checks address accidental inclusion of development metadata that a successful
local test run can overlook. This adds specific release checks; it does not
make every optional ReleaseKit validator mandatory or prohibit development
workspaces and replacements outside the selected committed candidate inputs.
Selecting only formatting and Markdown would change less release behavior, but
would leave these distribution mistakes to manual review. The selected upstream
checks avoid a second local implementation.
Host-normalized registered-tool execution is selected because the root module
already registers ReleaseKit and Go caches the executable. Retaining a persistent
native executable would provide a recovery entry point independent of a later
Go-driver invocation, but would retain executable-storage and bootstrap test
maintenance. Recovery instead uses the registered tool or ReleaseKit's reported
entry point. The credential directory remains ignored under either approach.
The implementation plan records the existing integration dependency-version
mismatch separately from validation of the current source. This infrastructure
sprint does not publish releases or select new library dependency versions.
Historical hosted-page backfilling and CI delivery remain separate future
choices, not prerequisites for the selected local page-only workflow.

View File

@@ -1,184 +0,0 @@
# ReleaseKit v0.9.0 adoption
Status: R1 and R2 remediation, ReleaseKit v0.9.0 adoption, and the completion
assertion repair are complete. The [completed implementation record](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/c8d659dfe5ae3e834b3aeabdfd0fe29b736dc05a/docs/roadmap/implementation.md)
records stage order, tasks, and validation evidence. This document owns the
target and selected choices; it does not authorize source or hosted publication.
## Purpose and authority
Adopt ReleaseKit v0.9.0's improved recovery behavior while preserving
CommandKit's existing release-tooling architecture. Update current dependency
and maintainer documentation together, and make the validation evidence clear
about existing local defects.
The [agent policy](../policy/agents.md),
[release policy](../policy/release.md#candidate-validation),
[testing policy](../policy/testing.md#upstream-integration-testing-ownership),
and [documentation policy](../policy/documentation.md) govern the work. Current
behavior remains owned by the [release procedure](../release.md), Go declarations,
and owning package tests until implementation updates them.
## Current baseline and review evidence
Reviewed CommandKit code baseline: `4d92ee4ee6a8403b22a4172e0f04face802b9499`.
The original upgrade review used `98e7e9933353cf43f2be4a96cb0436a0bce068d2`;
its findings and validation history are retained below.
- The root [module](../../go.mod) selects
`go.maximumdirect.net/releasekit v0.8.0` as a Go tool. Its selected version
serves core, ckconfig, and ckprompt operations.
- The [shared helper](../../scripts/releasekit) already normalizes the Go host,
disables workspace discovery, selects the local toolchain, retains read-only
metadata, and invokes the full executable package path. The
[legacy entry point](../../scripts/release) delegates publication to it.
- All three profiles already select the agreed generic file checks, forbidden
token path, and page-only Gitea destination. Source publication and hosted
delivery are separate operations. The local token directory remains ignored.
- [R1 and R2](remediation.md) are complete. Commit `7eac3fe` rejects invalid
legacy selectors before invoking Go; commit `4d92ee4` restores explicit,
optional POSIX-shell invocation in launcher tests. The remediation record owns
their acceptance evidence, including passing focused tests, shell syntax
checks, and profile assertions running when no shell is available.
- Both integrations still require older published CommandKit dependencies than
their current source needs. The completed [release-tooling implementation record](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/87eb84ccb462bf54e6250c8b287ac217efe90883/docs/roadmap/implementation.md#stage-5-verify-the-final-source-and-record-validation-boundaries)
distinguishes workspace-backed tests from independent release readiness.
The upstream v0.9.0 tag identifies commit
`2180aa868ec2caa95fc8279fd59d5f0ebac41bd4`. The tag and direct Go-module
resolution were verified. The default proxy initially returned v0.8.0 for
`@latest`; select v0.9.0 explicitly rather than changing repository-wide proxy
settings to work around that observation.
Upstream [release notes](https://forge.maximumdirect.net/base-layer/releasekit/src/tag/v0.9.0/docs/releases/v0.9.0.md)
describe host-safe launchers and generated recovery commands, full package-path
tool selection, preservation of wildcard and version-specific replacements in
API-only recovery, and safer manual hosted-page retries. Public Go APIs,
profiles, record formats, and the Go 1.26.0 consumer minimum remain unchanged.
The v0.8.0-to-v0.9.0 comparison also left upstream module requirements, checksums,
copyright, and license texts unchanged.
A disposable CommandKit upgrade changed only the selected ReleaseKit version in
`go.mod`. With Go 1.26.8, the root suite passed on a subsequent run, both
workspace-backed integration suites passed, and the real helper reported v0.9.0
under foreign target settings. No production publication or delivery was run.
A later disposable review at `7eac3fe` confirmed that downloading and tidying
v0.9.0 changed only the ReleaseKit requirement and its checksum entries. The
offline root suite and both workspace-backed integration suites passed with
Go 1.26.8, and the helper reported v0.9.0 under Windows/ARM64 target settings.
That snapshot excluded the then-uncommitted R2 change; it does not replace
validation of the eventual adoption commit.
The initial upgrade run failed in
[`TestCompletionCommandContext`](../../completion_context_test.go). The same
alias-order failure reproduced four times in 30 focused repetitions on the
unchanged v0.8.0 baseline. A single passing rerun therefore does not establish
stable validation. The test compares an internal alias slice in a fixed order;
commit `0acbbf4` deliberately changed completion to use unsorted lookup records
before this upgrade review.
## Targeted end state
### Tool adoption and retained architecture
The root module selects ReleaseKit v0.9.0 with authentic downloaded checksums.
Its existing tool declaration and root-owned selection remain. No integration
dependency version, library import path, Go minimum, or selected Go 1.26.8
maintainer toolchain changes as part of adoption.
Keep host normalization and full-package invocation in the existing helper:
upgrading a dependency does not update or replace local shell scripts. Keep one
module/profile mapping and the established wrapper forms. Retain the selected
profiles, token-file handling, credential-directory ignore rule, and separation
of source and hosted outcomes. No new validator, binary target, CI workflow,
publication channel, credential preflight gate, or runtime API is needed.
The new generated Go recovery commands complement the helper's existing host
handling. Recovery continues to use saved publication evidence and the upstream
record formats; CommandKit does not acquire a parallel recovery implementation.
### Release-code simplification review
The v0.9.0 review identified no additional local release implementation to remove.
The earlier simplification already delegated generic validation, publication,
delivery, and recovery to ReleaseKit. This release improves launcher examples
and generated recovery commands without adding validation options that replace
the remaining CommandKit code.
Retain the shared helper's module/profile selection and execution environment.
Host normalization must happen before Go compiles and runs ReleaseKit. Retain
the legacy publication wrapper to preserve the established maintainer interface.
The upstream multi-module launcher example adds a module map and temporary
metadata preparation for a different tool-selection model; replacing our
root-owned registered-tool helper with it would add machinery.
Retain the focused [wiring tests](../../scripts/release_test.go) for profile
selections, argument forwarding, environment, streams, status, and optional-shell
handling. ReleaseKit owns validation and recovery semantics; CommandKit's tests
continue to cover its own adapter. The adoption work therefore preserves the
release-code structure while updating the dependency and documentation, alongside
the separately planned completion assertion repair.
### Current documentation and history
Current release procedure, project-specific release policy, consumer Go-tool
selection summary, and root attribution describe v0.9.0. Upstream contract and
notice links use the corresponding immutable source revision. Record that the
host-safe launcher request has shipped upstream instead of describing it as
pending in the earlier feature roadmap.
Continue linking to upstream recovery and delivery contracts from the maintainer
procedure. Use ReleaseKit's generated recovery commands and keep local guidance
focused on CommandKit's entry points and selections.
Preserve the baseline and validation history in completed implementation and
review records. Do not mechanically replace every historical v0.8.0 reference.
Integration attribution describes its own selected published dependency graph;
update it only if that actual selection changes. No license-text replacement
is needed for the unchanged upstream texts.
### Completed remediation and remaining correctness work
R1 and R2 were delivered before the dependency upgrade, with their completion
markers and acceptance evidence in [remediation.md](remediation.md). Preserve
their selector rejection, legitimate raw helper forms, and optional-shell test
boundary during adoption; no further implementation of those findings remains.
The completion assertion compares the expected alias source/target records
without depending on their iteration order, while detecting missing, unexpected,
or duplicate records. Its routing and visible-completion assertions remain.
Production alias ordering and public contracts are unchanged. This correction
was completed as part of the maintenance scope, separate from R1/R2 remediation;
the implementation plan records the test changes and focused verification.
Ordinary releases retain their single selected-module publication suite and
separate hosted-delivery outcome. One-time upgrade diagnostics are maintenance
evidence, with no new recurring release gate. Credentials and live publication
are unnecessary for this adoption.
## Completion evidence
Adoption is demonstrated by authentic module metadata, the real helper reporting
v0.9.0 for the host, aligned current documentation and attribution, and recorded
validation of the final source. The reviewed upgrade introduced no other module
version change; verify rather than assume that result during implementation.
R1/R2 acceptance evidence is complete in the remediation record. Completion of
the remaining maintenance work also includes a stable, order-independent
completion assertion and validation that the adopted helper retains the
corrected launcher behavior. Independent integration dependency alignment
remains separate release-preparation work; this roadmap selects no future
versions or sums.
## Resolved choices
R1 and R2 were corrected in separate commits before adoption, resolving the
former OD-1 scheduling question. Their acceptance evidence remains in
[remediation.md](remediation.md).
The selected maintenance scope is the completion assertion repair, root tool
upgrade, aligned current documentation and attribution, and validation of the
resulting source. The older integration dependency-version mismatch remains
separate release-preparation work. The implementation plan defines execution
and commit sequencing.

View File

@@ -1,116 +0,0 @@
# Release tooling review findings
Status: R1 and R2 complete; no open findings. Review baseline:
`124a9fa57fb3f815bedbafe2d4819f6aa265ee7d`, covering the implementation commits
after `cb40f1d`. The findings below preserve the review evidence; completion
markers record the subsequent corrections. No implementation was changed during
the original review.
The selected architecture is substantially present: three declarative profiles,
one host-normalized registered-tool helper, separate publication and page-only
delivery, retained credential-directory ignoring, and focused wiring tests.
Both issues identified by the review have been addressed.
## R1 — P2: Reject raw helper command names as legacy module selectors
**Completed.** The legacy entry point now checks its three module selectors
before delegating to the helper, retaining the single profile/module mapping
there. Regression cases cover all three raw-command collisions and require
usage status 2 with no Go invocation. The new regression failed before the fix;
the complete focused scripts suite and shell syntax checks passed afterward
with Go 1.26.8, including valid legacy and direct raw-command forwarding.
**Location:** [scripts/release](../../scripts/release), lines 9–10, delegating to
the first-argument dispatch in [scripts/releasekit](../../scripts/releasekit),
lines 10–15. The invalid-form cases in
[scripts/release_test.go](../../scripts/release_test.go), lines 186–200, miss
these collisions.
The legacy interface is `scripts/release MODULE VERSION`, with three module
selectors. It now delegates its first argument directly into a helper that also
accepts `version`, `publication-attempts`, and `reconcile` as raw commands. These
invalid module names consequently bypass selector rejection and invoke Go.
For example, `./scripts/release version v1.2.3` reaches the tool as
`version publish v1.2.3`. With the real pinned tool it returns upstream's
"version takes no arguments" diagnostic after tool resolution. In a disposable
copy with a controlled Go boundary returning status 73, each of the three raw
command names returned 73 and invoked Go; an ordinary unknown selector returned
2 without invoking Go. Thus invalid legacy input depends on tool availability
and behavior instead of being rejected at the wrapper boundary.
**Required correction:** Make legacy module selection unambiguous before raw
command dispatch. Keep one profile/module mapping and the supported raw helper
forms; a small selector check is preferable to another routing framework.
Preserve upstream ownership of version and flag validation.
**Verification:** Add all three raw command names as invalid legacy selectors.
Each must return usage status 2 before any Go command runs. Valid legacy
publication and the helper's direct raw forms must retain their forwarding and
exit-status behavior. This restores the requested malformed-wrapper handling;
it adds no new accepted-version restriction.
## R2 — P2: Restore the tests' optional POSIX-shell execution boundary
**Completed.** Launcher execution tests now discover `sh`, invoke it explicitly
with the script path, and skip when it is unavailable. Profile assertions remain
independent of shell availability. A regression case removes the fixture
script's executable permission to verify interpreter invocation; it failed
before the correction and passes afterward. The focused scripts suite and shell
syntax checks passed with Go 1.26.8 on the available POSIX host. Running the
compiled test binary with an empty `PATH` confirmed that every launcher
execution case skips while all three profile assertions pass. Native Windows
execution was not tested and no new platform guarantee is introduced.
**Location:** [scripts/release_test.go](../../scripts/release_test.go), especially
`fixture.run` at line 63 and the launcher test setup.
The rewritten tests execute extensionless POSIX shell scripts directly through
`exec.Command`. They neither discover a shell nor skip launcher execution when
it is unavailable. The previous suite resolved `sh`, skipped shell-dependent
tests when it was missing, and passed the script as an argument to that shell.
Direct execution works on this review's Linux host, but relies on Unix script
execution semantics. Native Windows process creation does not supply a POSIX
interpreter for these files. Go's
[os/exec documentation](https://pkg.go.dev/os/exec#hdr-Overview) explicitly notes
that it does not invoke a system shell. Consequently the default repository
suite now fails on hosts without compatible script execution rather than
preserving the previous optional-shell behavior. This portability finding is
based on the changed execution path and platform implementation; a native
Windows test run was not available during review.
**Required correction:** Discover a POSIX shell for launcher execution tests,
invoke it explicitly with the script path, and skip those tests when the shell
is unavailable. Keep profile-only JSON assertions runnable without requiring a
shell. Preserve the reduced test architecture and existing argv/status checks.
No native Windows release launcher or additional supported-platform guarantee
is requested.
**Verification:** Confirm that launcher tests use the discovered interpreter
and handle its absence explicitly, while profile assertions still run. Rerun
the focused suite on the available POSIX host. Broader platform matrices are
not a new release gate; this correction removes an incidental environment
restriction introduced by the test rewrite.
## Review evidence and limits
- Go 1.26.8 offline root tests and the temporary-workspace suites for both
integrations passed. Both production scripts passed shell syntax checks.
- The real helper reported ReleaseKit v0.8.0 under foreign target settings.
- ReleaseKit checks passed for a disposable committed source snapshot using
the actual profiles: ordinary root validation, then both integration checks
with the implementation plan's temporary workspace-backed test process.
Selected file checks continued to inspect the committed declarations.
- Local documentation destinations and applicable Markdown fragments passed
inspection. The credential file remained ignored and its contents were not
read. No real publication or hosted-delivery operation was invoked.
- Library code and dependency declarations were unchanged by the sprint. The
documented integration dependency-version mismatch remains a pre-existing
independent-release-readiness limitation; the workspace-backed results do
not establish that published dependencies are aligned.
No additional architecture gaps or unrequested runtime/public-API guarantees
were identified. The selected file checks and forbidden token path match the
finalized scope. R2 concerns the default tests' execution environment rather
than a new production validator or library contract.

View File

@@ -1,890 +0,0 @@
# Output policy and invocation reporting
Status: behavioral decisions selected through OD-4. Message categories, all four
output modes, uniform treatment of help and version, message prefixes, and
OD-1's flag grammar and scope (including independent Progress control) are selected.
OD-2's unified reporting direction, the terminal failure model, and application
ownership of domain-specific error decisions are selected.
OD-3's failure presentation and retention boundaries are selected.
OD-4's delivery lifecycle, write-failure handling, and quiet-mode rendering are
selected. The consumer API signatures, presentation-result model, provider
diagnostic boundary, fallback explanations, and rendering callback boundary
below are also selected. Architectural preparation is complete:
[ADR 0014](../adr/0014-invocation-reporting.md) is accepted, and the
[architecture policy](../policy/architecture.md) is aligned. Implementation and
final validation are complete; Stages 1–20 are complete. The [implementation
plan](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/230187296fae5253cb1773f16c2f1d1a39499ed0/docs/roadmap/implementation.md) records the stage order and validation evidence.
## Purpose
Provide consistent handling of `--quiet`, `--verbose`, `--debug`, and
`--progress` across application commands and optional integrations. Reduce
repeated output classification, diagnostic presentation, and delivery bookkeeping
where CommandKit can provide a meaningful benefit.
Ordinary Go writers and formatting functions already make application output
straightforward. Design reporting conveniences around the shared output policy
and invocation lifecycle. Applications continue to own their report content,
domain behavior, and operation outcomes.
## CLI presentation and error-handling ownership
Applications decide how to respond to domain-specific errors: recover, retry,
continue silently, report a Warning, or end the invocation with a failure.
CommandKit may provide shared helpers for carrying, inspecting, and presenting
errors while leaving those decisions with the application. A Go `error` alone
does not determine a CLI message category. Warnings may describe issues with or
without an underlying Go error, and handling an error does not require emitting
a message.
CommandKit handles the presentation the application requests: message category,
visibility, formatting, timing, and delivery. For terminal application failures,
the application supplies a non-zero outcome and an operator-facing explanation.
It may also supply an optional Go error cause for the execution caller to inspect.
CommandKit preserves that cause and its existing `errors.Is`/`errors.As` behavior.
Supplying a cause alone does not select an outcome, generate an operator-facing
message, or change visibility.
Applications can explicitly present `err.Error()` as a Warning through the
ordinary message API, illustrated by `report.Warn(err.Error())`. Defer a separate
`WarnError` convenience helper pending demonstrated value; this conversion
already requires little code.
CommandKit retains responsibility for handling its own errors, including
routing, parsing, and reporting failures. Integration-owned commands similarly
adapt their upstream APIs into deliberate command outcomes and presentations.
Ordinary Go error returns from CommandKit APIs and writers remain available where
appropriate.
## Current baseline
[`Streams` and `Invocation`](../../identity.go) expose caller-selected output
and diagnostic writers. [`Outcome` and `Diagnostic`](../../outcome.go) carry
operation causes, diagnostic writers, and delivery state. Handlers can return
`Failure(err)` or attach diagnostics to an outcome for presentation after the
handler returns. `MessageDiagnostic` currently writes the supplied message with
a newline; severity does not add a label automatically.
CommandKit coordinates final diagnostic delivery and retains delivery failures
separately from the operation cause. Custom handlers can also write directly to
the supplied streams and retain responsibility for their output. The existing
outcome and channel ownership decisions are recorded in
[ADR 0005](../adr/0005-invocation-outcomes.md).
Integration adapters already translate upstream diagnostic records into
CommandKit diagnostics while preserving native rendering and source context.
Some adapters wrap multiple records in one diagnostic writer and classify the
batch by its highest severity. This matters when designing selective visibility.
There is currently no shared verbosity policy or invocation reporter API.
## Selected output policy
Distinguish substantive application output from diagnostic message categories:
- **Substantive output (Normal):** The requested report or command result.
- **Error:** Terminal presentation of an unrecoverable invocation failure,
resulting in a non-zero status code.
- **Warning:** Something went wrong or deserves attention, but the invocation
can still succeed. The application chooses whether to report such an issue.
- **Info:** Informational context and useful explanation, such as which profile
was selected or a final operation summary.
- **Progress:** Streaming updates about work underway, such as items processed.
- **Debug:** Troubleshooting detail, including intermediate decisions and
execution-path information.
Debug mode adds troubleshooting detail to verbose output:
| Output category | `--quiet` | No flag | `--verbose` | `--debug` |
| --- | --- | --- | --- | --- |
| Substantive application output | Suppressed | Shown | Shown | Shown |
| Error | Shown | Shown | Shown | Shown |
| Warning | Suppressed | Suppressed | Shown | Shown |
| Info | Suppressed | Suppressed | Shown | Shown |
| Debug | Suppressed | Suppressed | Suppressed | Shown |
| Progress, without an explicit setting | Suppressed | Suppressed | Suppressed | Shown |
Info belongs in verbose mode; debug includes verbose messages. Normal output
remains a substantive result category, separate from diagnostic severity.
The category of a message determines its visibility; stream destination alone
cannot distinguish a warning from an error.
Progress is distinct from Info and is controlled separately as specified below.
Applications and integrations are not required to generate messages in any
category. Enabling a category adds no output when there are no messages in it;
other output remains subject to its own visibility rules.
Error presentation follows the terminal failure model below. An application
may report a recovered problem as a Warning or continue without a message.
Apply the same policy to help and version. There are no command-specific
exceptions for successful output:
```text
app help --quiet # succeeds silently
app version --quiet # succeeds silently
```
Usage and operational failures still report their errors in quiet mode.
The feature must distinguish successful output suppression from failure
diagnostic delivery.
### Message prefixes and channels
Substantive output uses the result destination, conventionally stdout. Warning,
Info, Progress, Debug, and terminal explanations use the diagnostic destination,
conventionally stderr. Callers supply both destinations.
Use `WARN:`, `DEBUG:`, and `ERROR:` prefixes for simple warning, debug, and error
messages respectively. Substantive output and Info messages have no prefix.
For multiline diagnostics, prefix the opening line and preserve the structured
detail. Integration adapters coordinate native diagnostic rendering to avoid
duplicate severity labels or loss of upstream formatting.
Usage guidance belongs to the terminal usage explanation and remains visible in
quiet mode. Successful help output follows the substantive-output policy.
Initially, Progress messages are unprefixed streamed lines. The diagnostic
destination keeps them separate from substantive output such as JSON or YAML.
Terminal redraw and progress-bar behavior are outside this implementation.
### Terminal failure model
During execution, applications can produce substantive output, Progress, Info,
Debug, and Warnings. An unrecoverable failure ends the handler by returning a
failed outcome. There is no standalone `report.Error(...)` method. Detailed
error explanations accompany the returned failure and are presented as part of
that terminal step.
CommandKit finalizes reporting, presents the terminal failure explanation, and
returns the non-zero outcome to the application entry point. It accepts no
further normal output for that invocation after terminal error presentation.
The same boundary applies to core routing and parsing failures that prevent
handler execution. The application entry point retains ownership of process
exit; CommandKit ends the invocation and returns its outcome.
For example, if an export operation fails and the application decides it cannot
continue, the handler returns a non-zero outcome with an explicit explanation
such as "Unable to export the requested report." The application can choose to
include detail from a Go error in that explanation. CommandKit presents the
supplied explanation without automatically deriving text from an application
error. The selected API below provides failure constructors for this purpose.
When a failed outcome supplies no explanation, CommandKit supplies the generic
fallback defined under terminal outcomes below.
Output streamed before the failure remains visible according to its selected
mode; terminal failure does not withdraw earlier bytes. Deferred messages,
cleanup writes, and delivery failures must respect the end of normal reporting.
Their finalization and late-write behavior follow OD-4 below. A terminal
failure may have multiple explanatory records, all belonging to that failure.
### Failure explanations and presentation ownership (OD-3)
Make a non-zero outcome with an explicit message the convenient failure path.
Support structured explanations, including multiple records, through the same
presentation facility for richer application and integration reports. Neither
form requires an underlying Go error.
The application or integration explicitly identifies the context belonging to
a terminal explanation. Present that context with the failure in every mode,
including quiet. Independent warnings and other messages retain their own
visibility rules. For mixed upstream batches, adapters preserve useful native
rendering while distinguishing terminal explanatory context from independent
messages. Core does not infer those relationships from error chains or text.
Assign each submitted message one presentation path. Returning a terminal
explanation schedules its presentation once; core does not append another
message derived from an application error cause. Native renderers can supply
already-labelled content without receiving duplicate prefixes. Avoid text-based
deduplication: separate events may legitimately produce identical messages.
Constructing a failure result attaches explanatory content without emitting or
independently scheduling an Error. Returning that result hands the explanation
to CommandKit for terminal delivery through the invocation's reporting facility.
The selected API expresses this distinction for both simple and structured
explanations. For example:
| Handler action | Presentation |
| --- | --- |
| Construct a non-zero result with "Unable to export the requested report," then return it. | Construction emits nothing. CommandKit presents the message once during terminal delivery. |
| Return a failure with native diagnostic records and their source context, submitting independent warnings separately. | CommandKit presents the structured terminal explanation in every mode. Independent warnings follow their own visibility policy. |
### Retention boundary (OD-3)
CommandKit provides invocation-scoped CLI presentation. Logging and message
history collection remain application concerns. Keep terminal explanations and
any supplied cause with the returned outcome, and keep the temporary state needed
for deferred delivery, parser events awaiting mode selection, and delivery
bookkeeping. The reporting facility does not collect a transcript or retain other
emitted or suppressed messages for later inspection. No optional message-history
facility is planned for this implementation.
## Selected flag grammar and scope
### Mode selection and resets
Model quiet, ordinary, verbose, and debug as mutually exclusive presentation
modes. Each invocation starts in ordinary mode. Process flag occurrences in
invocation order; the last enabled mode wins. Repetition does not increment
verbosity, and debug is selected explicitly.
Accept presence and the existing explicit boolean syntax, including
`--quiet=true` and `--quiet=false`. A false occurrence clears its mode only if
that mode is currently selected, returning to ordinary mode. It does not restore
an earlier selection or disable a different mode. Every occurrence is converted;
invalid boolean values remain usage errors even when a later flag would replace
the selection.
| Flag sequence | Final mode |
| --- | --- |
| `--quiet --verbose` | Verbose |
| `--debug --quiet` | Quiet |
| `--verbose --verbose` | Verbose |
| `--verbose --verbose=false` | Ordinary |
| `--quiet --verbose=false` | Quiet |
| `--verbose --quiet --quiet=false` | Ordinary |
| `--quiet=false --debug` | Debug |
Keep verbosity a presentation choice: quiet preserves the requested operation's
semantics and rendering failures. Delivery failures depend on the output actually
requested; suppressing a destination write can avoid a delivery failure. Apply
the policy through CommandKit-managed reporting and streams. Output written
directly to process streams or external destinations remains application-owned.
### Progress selection
`--progress` enables Progress without enabling Info, Warning, or Debug messages.
`--progress=false` explicitly disables it. Accept the existing boolean syntax;
repeated canonical or aliased progress settings use the last value. Retain
whether a setting was explicitly supplied, separately from its boolean value.
Determine effective Progress visibility from the final presentation mode and
explicit progress setting:
1. Quiet mode always suppresses Progress.
2. Otherwise, an explicit progress setting controls visibility.
3. Without an explicit setting, debug enables Progress; ordinary and verbose
modes leave it disabled.
An explicit progress setting overrides the mode's default regardless of its
position relative to mode flags. It does not select or reset the presentation
mode. Debug includes verbose detail and defaults Progress on; verbose alone
does not enable Progress.
| Flags | Result |
| --- | --- |
| `--verbose` | Warning and Info, without Progress. |
| `--verbose --progress` | Warning, Info, and Progress. |
| `--debug` | Warning, Info, Debug, and Progress. |
| `--debug --progress=false` or `--progress=false --debug` | Debug detail without Progress. |
| `--quiet --progress` or `--progress --quiet` | Only failure diagnostics. |
| `--debug --verbose` | Verbose detail without Progress, absent an explicit progress setting. |
Passing `--progress` to an application that emits no Progress messages adds no
visible output. It does not suppress that application's substantive results or
require a progress implementation.
### Placement, availability, and aliases
Core supplies `--quiet`, `--verbose`, `--debug`, and `--progress` for application
commands, integrations, the root, canonical groups, and alias-only discovery help.
Consumers do not register them individually. Preserve the existing placement
grammar: flags follow the complete command path and may be interspersed with
operands after that path. For example:
```text
app config show --quiet
app export destination --verbose
app help config show --quiet
app version --quiet
```
Options before or between command-path components are outside this grammar.
Tokens consumed as values and tokens following `--` retain their ordinary
meaning.
Reserve the four long spellings. Applications may add `-q`, `-v`, or `-d`
through the alias API; no built-in short spellings are selected. An application
flag declaration that defines a reserved long name fails installation with a
collision error. Aliases may target these flags but may not overload canonical
spellings. Apply the existing collision and atomic installation rules, including
later registration.
Aliased and canonical occurrences participate in the same ordering. With
`-q → --quiet`, `app export --verbose -q` selects quiet mode. Document the newly
reserved canonical names as a compatibility change.
### Routing and parsing failures
Determine the mode through the normal parser before execution or successful
help/report presentation. Do not add a separate verbosity scanner. This keeps
token interpretation consistent and avoids reinterpreting values or operands.
Routing and parsing failures report their errors regardless of verbosity. For
example, `app export --quiet --unknown` still reports the unknown flag. Error
visibility follows the selected output policy independently of how the mode is
determined; avoiding a separate scanner is an implementation simplification.
## CommandKit's own diagnostics
Use the shared reporting system for CommandKit's invocation diagnostics so
parser, integration, and application messages follow the same visibility and
delivery policy. Installation errors occur before an invocation exists and
continue to be returned to the application without automatic printing.
Supported repetition and last-occurrence overrides are routine decisions.
Repeated singleton flags, including `--verbose --verbose`, may produce debug
detail; repetition alone does not warrant a warning. Repeatable flags retain
their intended accumulation behavior. Add warnings for concrete concerns an
operator can act on, such as a deprecated spelling if deprecation is introduced.
Unknown flags, invalid values, and invalid command paths remain invocation
failures with error diagnostics.
Start debug instrumentation with routing, alias resolution, verbosity selection,
and handler completion. Other useful candidates can be added where they
explain a consequential decision:
| Event | Useful detail |
| --- | --- |
| Command-alias resolution | Entered path, intermediate redirects, and final canonical path. |
| Flag-alias resolution | Entered spelling and canonical flag name. |
| Scoped alias selection | A command-scoped mapping took precedence over an application-wide mapping. |
| Verbosity selection | A later occurrence selected a different mode; for `--verbose --debug`, debug wins because it appears last. |
| Help-route selection | A canonical command or alias-only discovery context was selected. |
| Repeated singleton flag | A later occurrence replaced an earlier occurrence, without printing their values. |
| Input preparation | Canonical input validation completed and the handler is starting. |
| Handler completion | Canonical command and the handler's returned classification and status, before final delivery. |
The handler-completion event describes the handler's result. The final returned
invocation outcome also includes delivery failures and remains authoritative.
Present this event before terminal failure presentation; do not emit a completion
message afterward.
Use the final selected mode to present earlier parser decisions. Collect the
relevant events during parsing and apply visibility after mode selection:
`--verbose --debug` shows the selection explanation, while `--debug --quiet`
suppresses it. This collection of parser events does not require whole-run
output buffering.
Debug messages should explain operator-relevant decisions independently of
incidental implementation steps. Use command paths, flag names, and decision
descriptions rather than raw argument or configuration-value dumps. Upstream
libraries retain ownership of their diagnostic detail and disclosure rules.
## Selected delivery lifecycle and quiet execution (OD-4)
### Ordering and finalization
Use synchronous delivery. Present immediate messages and substantive writes in
submission order. After the handler returns:
1. Deliver eligible deferred messages in queue order, without sorting by severity.
2. Finish ordinary reporting, including any handler-completion debug message and
flushing any CommandKit-owned buffers for ordinary output. Account for failures
discovered during this delivery before preparing the terminal explanation.
3. Present any terminal failure explanation and complete its delivery, including
any CommandKit-owned buffering of that explanation.
4. Close the invocation's reporting facility and return the final outcome.
Deferred messages remain eligible when the operation fails, subject to verbosity.
Deferral controls timing; applications queue success summaries only after the
relevant operation succeeds. Ordering governs CommandKit's writes; separate
stdout and stderr destinations may be buffered or displayed differently by the
caller.
CommandKit finalizes automatically. Consumers do not need a separate flush or
finish call. Handler cleanup, including Go `defer` functions, runs before the
handler returns and participates in normal reporting. Applications complete any
reporting goroutines before returning. After finalization, writes and message
submissions return a closed-reporter error and emit nothing.
Reporter operations within an invocation are safe to call concurrently.
Serialize access to reporting state and delivery so individual messages do not
interleave. Concurrent submissions have no predetermined order; deferred messages
follow the order in which they enter the queue. Multiple writes composing one
application report still require application coordination if they must remain
together. This contract covers operations through one invocation's reporter;
coordination with writes outside that reporter remains caller-owned.
Finalize any buffers owned by CommandKit. Caller-supplied writers remain
caller-owned; CommandKit does not automatically close them or invoke arbitrary
flushing interfaces.
Visibility applies when output is emitted. A verbose or debug run may show
informational detail before subsequently failing. Progress and substantive
output may likewise be partially delivered before a failure. Previously written
bytes cannot be withdrawn. Whole-run buffering solely to make failed runs
contain only error output is outside this implementation.
Keep errors understandable in every mode, including source locations and
related explanatory context. Unrelated warnings and informational messages
remain subject to the selected mode. Apply the selected retention boundary
without accumulating a message history.
### Rendering failures, write failures, and execution flow
Distinguish who invokes a renderer and whether its destination failed:
| Failure source | Responsibility and handling |
| --- | --- |
| An encoder or renderer called directly by the application | The application handles its returned error and chooses the outcome and explanation. |
| A rendering callback invoked by CommandKit | CommandKit handles its returned error as a presentation failure and reflects it in the invocation outcome. |
| A destination rejects a write | CommandKit records the delivery failure and stops further writes to that destination. |
A rendering failure may occur before any write or while the destination remains
healthy. It does not by itself establish a failed destination. CommandKit can
explain its rendering failure through a working diagnostic destination. These
responsibilities also apply when rendering into a discard destination under
quiet mode.
Failure to deliver requested output makes an otherwise successful invocation
unsuccessful. Apply this rule to substantive output, Warning, Info, Progress,
Debug, and terminal explanations. Intentional suppression requests no delivery
and is not a delivery failure.
Standard writer operations return write errors normally. Immediate reporting
methods also expose delivery errors. CommandKit records bounded delivery-failure
information so an ignored write error cannot yield a successful invocation.
Deferred delivery failures appear in the final outcome.
Apply ADR 0005's outcome rules when presentation fails. For standard outcomes,
return operational failure with status `1`, including when the original outcome
was a usage failure. Preserve the original explanation and any supplied cause;
record the presentation failure separately in `Presentation`.
Application-defined domain outcomes retain their domain classification. Preserve
an already non-zero domain status. If the domain status was `0`, change the
suggested status to `1` so presentation failure cannot yield a successful
invocation. The application entry point retains control of process exit.
| Outcome before presentation fails | Final outcome |
| --- | --- |
| Success, status `0` | Operational failure, status `1`. |
| Usage failure, status `2` | Operational failure, status `1`. |
| Operational failure, status `1` | Operational failure, status `1`. |
| Domain outcome, non-zero status | Same domain classification and status. |
| Domain outcome, status `0` | Same domain classification, status `1`. |
These rules apply to both destination-write failures and rendering failures from
callbacks invoked by CommandKit. Reporting failures remain CommandKit's own
error-handling responsibility.
Stop writing to a destination after it fails. A substantive-output failure can
receive a terminal explanation through the diagnostic destination if it still
works. If diagnostic delivery fails, return that information to the caller.
Do not redirect diagnostics into substantive output or recursively attempt to
report the reporting failure.
A failed write does not stop application code. The handler can inspect the error
and return immediately; otherwise, CommandKit records the failure and finalizes
when the handler returns. CommandKit does not use process exit or panic to
interrupt execution. Terminal failure presentation ends normal output.
### Quiet-mode rendering and suppression
Initially, run substantive report generation normally with its managed writer
discarding output in quiet mode. This preserves report generation, validation,
and serialization failures while changing presentation. For example, a report
encoder failure still fails the invocation under `--quiet`; the handler remains
responsible for handling errors returned by an encoder it calls.
Simple suppressed messages need no formatting or destination writes. For richer
report renderers, initially render into a discard destination. Rendering may be
skipped later where doing so preserves the command's operation and failure
semantics.
Treat suppression as intentional suppression, never as failed or incomplete
delivery. Do not retain suppressed content. Delivery needs only the deferred
queue, reporting lifecycle state, and bounded delivery-failure information,
alongside the temporary parser state and returned explanations described above.
## Selected consumer reporting API
Provide one reporting facility per invocation for substantive output, immediate
messages, and end-of-run diagnostics. The consumer API is chosen for its intended
long-term design; with no current consumers, preserving existing API shapes
carries no weight in this decision.
The following declarations specify the target API and are not yet implemented.
The reporter owns presentation: output visibility, progress settings, prefixes,
ordering, and delivery bookkeeping. Its views share invocation state, allowing
already-emitted messages, deferred messages, suppression, and write failures to
be coordinated in one place. Helpers may receive the reporter or a narrow
application-defined interface. Domain libraries can continue returning
diagnostic data for the handler to submit, independently of CommandKit.
Use the reporter as the common diagnostic submission path, replacing the earlier
proposal to split submission between a reporter and `Outcome.Diagnostics`.
Terminal explanations accompany returned outcomes and are delivered through the
same facility. Caller-selected destinations remain at `App.Execute`; handlers
use `Invocation.Reporter()` in place of raw invocation streams.
### Invocation reporter and structured detail
```go
func (inv Invocation) Reporter() *Reporter
type Reporter struct {
// Private invocation state.
}
func (r *Reporter) Output() io.Writer
func (r *Reporter) RenderOutput(write func(io.Writer) error) error
func (r *Reporter) Info(message string, details ...Detail) error
func (r *Reporter) Warn(message string, details ...Detail) error
func (r *Reporter) Debug(message string, details ...Detail) error
func (r *Reporter) Progress(message string, details ...Detail) error
func (r *Reporter) AtEnd() DeferredReporter
type Detail func(io.Writer) error
```
`Output()` supports ordinary writers and encoders. `RenderOutput()` coordinates
an entire rendering operation, captures its rendering error, and prevents other
reporter operations from interleaving with it. Rendering callbacks write through
their supplied writer. Each complete categorized message, including its details,
is likewise coordinated as one reporter operation.
Rendering callbacks use their supplied writer only during the callback. They may
call rendering helpers that write through it. While a callback is running, it
must not recursively submit messages, queue deferred messages, start another
rendering operation, or write through a separate `Output()` handle belonging to
the same invocation's reporter. Make additional reporter submissions before or
after the callback. A callback must also avoid waiting for another goroutine to
submit through that reporter, since the submission may be waiting for the
callback to finish.
This boundary applies to `RenderOutput`, immediate and deferred message details,
and terminal explanation details. Ordinary concurrent reporter calls remain
supported and serialized. Keeping each rendering operation together motivates
the boundary; it does not prescribe a lock implementation or promise runtime
detection of recursive calls.
Message strings are literal; consumers use `fmt.Sprintf` when formatting is
needed. Initially omit parallel `Infof`, `Warnf`, and similar methods. CommandKit
formats the opening message with its category's prefix and a terminating newline.
Optional details follow in order with their formatting preserved. An empty
opening message produces no heading, prefix, or blank opening line, supporting
native renderers that supply a complete, labelled diagnostic:
```go
_ = report.Warn("Using the default profile")
_ = report.Warn("Some configuration settings need attention.", writeNativeWarnings)
_ = report.Warn("", writeNativeWarnings)
```
These are alternative calls; `writeNativeWarnings` represents an
application- or integration-supplied function with signature
`func(io.Writer) error`. No generic logging method or severity enumeration is
needed to submit messages. Applications need not check verbosity before reporting
or implement categories they do not use.
Immediate methods return rendering or delivery errors, and CommandKit records
the corresponding presentation failure even if the caller ignores the error.
A caller can inspect the returned error and end its handler immediately.
### Deferred messages
```go
type DeferredReporter struct {
// Private reference to the same invocation state.
}
func (r DeferredReporter) Info(message string, details ...Detail) error
func (r DeferredReporter) Warn(message string, details ...Detail) error
func (r DeferredReporter) Debug(message string, details ...Detail) error
```
For example:
```go
_ = report.AtEnd().Warn("Three optional files were unavailable.")
```
Copies of the deferred view share invocation state. Its methods return submission
errors, such as attempting to use a closed reporter. Errors encountered during
deferred rendering appear in the final invocation outcome. Progress remains an
immediate streaming operation; substantive output uses the main reporter.
Copy callback lists when accepting a submission. Callers retain ownership of
data captured by rendering callbacks and keep that data valid until presentation.
### Terminal outcomes
Outcome construction requires no invocation state and emits nothing:
```go
type Explanation struct {
Message string
Details []Detail
}
type Outcome struct {
Kind OutcomeKind
Status int
Domain string
Cause error
Explanation Explanation
Presentation PresentationResult
}
func Success() Outcome
func Failure(message string, details ...Detail) Outcome
func Usage(message string, details ...Detail) Outcome
func Domain(name string, status int) Outcome
func (o Outcome) WithExplanation(message string, details ...Detail) Outcome
func (o Outcome) WithCause(cause error) Outcome
```
`Success` returns a successful outcome with status `0`. `Failure` produces an
operational failure with status `1`; `Usage` produces a usage failure with status
`2`. `Domain` constructs an outcome with the supplied application-defined
classification and status. Subsequent presentation failures follow OD-4's rules:
domain classification is preserved, and a zero status becomes `1` if presentation
fails.
`OutcomeKind` retains the success, usage, operational, and domain classifications.
`WithExplanation` returns a modified outcome value and replaces its terminal
explanation. Explanations belong to failed outcomes; successful summaries use
Info or substantive output. Ordinary application failure, native failure
rendering, and a custom domain failure respectively can be expressed as:
```go
return commandkit.Failure("Unable to export the requested report.")
```
```go
return commandkit.Failure("", writeNativeFailure)
```
```go
return commandkit.Domain("validation", 3).
WithExplanation("The document failed validation.", writeFindings)
```
The named renderers above have signature `func(io.Writer) error` and preserve the
owning library's safe diagnostic formatting. They run during terminal delivery,
not when constructing the outcome.
Before terminal delivery, supply the explanation `Command failed.` when a
non-zero outcome has an empty `Explanation.Message` and no non-nil callback in
`Explanation.Details`. This produces `ERROR: Command failed.` in every mode,
including quiet, and becomes the explanation retained in the returned outcome.
Adding the fallback leaves classification, status, and cause unchanged;
presentation failures still follow OD-4's rules. The fallback is based on the
failed outcome and absent explanation, without deriving text from `Cause.Error()`.
An empty heading with a non-nil detail renderer is an explicitly supplied
explanation. Trust that renderer to provide its content. Do not buffer or replay
it to decide whether the explanation is sufficiently informative, and do not
add the fallback merely because it writes no bytes.
| Returned outcome | Terminal presentation |
| --- | --- |
| `Failure("")` | Generic fallback, operational classification and status `1`. |
| `Domain("validation", 3)` | Generic fallback, domain classification and status `3`. |
| `Failure("").WithCause(err)` | Generic fallback; the supplied cause remains available without automatic rendering. |
| `Failure("", writeNativeFailure)` | Native explanation without an added generic heading. |
`Cause` is optional. `WithCause` returns a modified outcome value, replacing the
cause without changing the classification, status, or explanation. For example:
```go
return commandkit.Failure("Unable to open the configuration file.").
WithCause(err)
```
CommandKit presents the supplied explanation and carries the original error
through to the execution caller, which can inspect it with expressions such as
`errors.Is(outcome.Cause, fs.ErrNotExist)`. It does not automatically present
`Cause.Error()` or use the cause to decide success, failure, or message category.
Providing a cause is independent of choosing the terminal explanation; neither
is inferred from the other. Warnings do not require supplying or retaining causes.
Outcome processing preserves supplied causes, including their existing Go error
identity, when a subsequent rendering or delivery failure occurs. Presentation
failures are recorded separately in `Presentation`. The returned outcome retains
its terminal explanation and optional cause, with other retention limited to the
presentation needs described above.
### Provider-returned diagnostic data
Providers return values and diagnostic data without needing a reporter parameter.
Retain `Acquired[T]` with separate fields for independent messages and failure
explanations:
```go
type Acquired[T any] struct {
Value T
Diagnostics []Diagnostic
Explanation Explanation
}
```
`Diagnostics` carries independent, nonterminal messages. `Diagnostic` and
`Severity` remain passive data types for those messages, with each diagnostic
carrying its `Severity` and a `Write func(io.Writer) error` renderer. Adapt their
contracts to the nonterminal reporting categories. Integrations submit the
records through the corresponding reporter methods, including the deferred view
where appropriate. Removing `Outcome.Diagnostics` removes a submission path;
provider-returned records remain available for the integration to submit.
`Explanation` carries context to attach when the integration returns a failed
outcome. Terminal Error content uses this field rather than an independently
submitted diagnostic. An upstream warning that helps explain the failure may
belong in the explanation; unrelated warnings remain in `Diagnostics` and follow
their ordinary visibility rules. Adapters preserve this distinction and useful
native formatting when translating mixed upstream batches, instead of assigning
the whole batch its highest severity.
Providers may return acquired data alongside a Go error. The owning integration
chooses the command outcome, attaches any terminal explanation, and decides which
cause to preserve. Returning diagnostic data does not itself emit messages or
select the invocation outcome. Diagnostic renderers follow the same callback
boundary as other details when invoked through the reporter.
### Presentation observations and closed reporters
```go
type PresentationResult struct {
Output DeliveryResult
Diagnostics DeliveryResult
}
type DeliveryResult struct {
State DeliveryState
RenderError error
WriteError error
}
type DeliveryState uint8
const (
DeliveryNone DeliveryState = iota
DeliverySuppressed
DeliveryComplete
DeliveryIncomplete
)
var ErrReporterClosed = errors.New("reporter is closed")
```
| State | Meaning |
| --- | --- |
| `DeliveryNone` | No content was submitted for this destination. |
| `DeliverySuppressed` | All submitted content was intentionally suppressed, and rendering succeeded. |
| `DeliveryComplete` | All requested delivery completed; other content may have been suppressed. |
| `DeliveryIncomplete` | Rendering or destination writing failed. |
Retain the first rendering failure and first destination-write failure for each
destination. This provides bounded evidence without collecting an event history.
`Outcome.Presentation` is populated by CommandKit during execution and
finalization; it is observation data, not another submission API.
The invocation supplies the reporter. Consumers neither construct nor finalize
it. After closure, message submissions and writer operations return
`ErrReporterClosed`, including in quiet mode. The initial API needs no explicit
flush method, transcript accessor, or verbosity checks. Error-conversion helpers
are deferred pending demonstrated value.
### Application example
This target-API handler uses an application-owned `exportData` operation that
returns a JSON-encodable result with a `UsedDefaults` field:
```go
func runExport(inv commandkit.Invocation) commandkit.Outcome {
report := inv.Reporter()
result, err := exportData(inv.Context)
if err != nil {
return commandkit.Failure("Unable to export the requested data.")
}
if result.UsedDefaults {
_ = report.Warn("Some optional settings used defaults.")
}
if err := report.RenderOutput(func(w io.Writer) error {
return json.NewEncoder(w).Encode(result)
}); err != nil {
return commandkit.Failure("Unable to present the export result.")
}
_ = report.AtEnd().Info("Export completed.")
return commandkit.Success()
}
```
Ignored immediate reporting errors are still recorded and can make the final
invocation unsuccessful. The renderer runs in quiet mode against the discard
destination, preserving serialization failures. The deferred summary follows
the selected verbosity policy and describes the completed export operation;
the final invocation outcome also accounts for presentation failures.
## Upstream diagnostic ownership
ConfigKit and PromptKit remain authoritative for their diagnostic semantics,
severity, source context, and disclosure rules. Integration adapters expose
those capabilities through the shared CommandKit presentation policy.
Support consistent reporting of ConfigKit errors and warnings as its adopted
public API evolves. This roadmap does not assume that a future warning API
already exists or prescribe its upstream shape. Integration-owned commands
select the outcome and submit useful native diagnostic detail for presentation.
Core presents those records under the shared policy. Applications using upstream
libraries directly choose how to respond to domain-specific errors and which
outcomes and messages to supply, as for their other operations.
An integration may receive both a useful substantive report and an operation
failure, as PromptKit smoke testing supports. Preserve the report under the
ordinary substantive-output policy, then present the terminal explanation.
Quiet mode suppresses the report while retaining the explanation. Having a
report does not imply a successful outcome, and returning a failure does not
discard substantive output already prepared for delivery before the terminal step.
Debug output follows the same upstream disclosure and safe-rendering contracts.
Enabling debug does not authorize secret disclosure or bypass owning-library
rules.
## Delivery scope
The completed feature includes the selected reporter and outcome API, ordered
mode/progress selection through normal parsing, shared core presentation, and
both optional integration migrations. Passive provider diagnostics distinguish
independent messages from terminal explanations while preserving native evidence
and rendering. Existing command behavior remains unchanged outside the selected
reporting policy.
Maintained consumers demonstrate ordinary output, immediate/deferred messages,
silent recovery, explicit warnings, failure after partial output, optional cause
transport, and structured terminal explanations. Integration consumers include
native diagnostic context and useful reports returned alongside failures.
Behavioral coverage establishes CommandKit-owned parsing, visibility, ordering,
concurrent reporter coordination, closure, presentation observations, and failure
handling. It preserves existing alias, input/state, discovery, native-report,
and upstream-adapter protections without testing upstream domain correctness.
GoDoc, CLI documentation, consumer guidance, internal documentation, and release
notes describe the implemented boundary under their canonical ownership.
The [implementation plan](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/230187296fae5253cb1773f16c2f1d1a39499ed0/docs/roadmap/implementation.md) defines the concrete migration
steps, focused validation, and completion criteria. This roadmap defines the
target behavior and selected policy choices.
## Architectural preparation completed
- [x] Review and accept [ADR 0014](../adr/0014-invocation-reporting.md) as an
extension for shared reporting, retaining ADR 0005's outcome and supplied-error
preservation decisions.
- [x] Align the [architecture policy](../policy/architecture.md) with the reporting
ownership model, including application ownership of domain-specific error
decisions and scope for useful shared error helpers.
OD-1 through OD-4 and the consumer API signatures are selected. The target end
state includes maintained, executable consumers using those APIs; the
[implementation plan](https://forge.maximumdirect.net/base-layer/commandkit/src/commit/230187296fae5253cb1773f16c2f1d1a39499ed0/docs/roadmap/implementation.md) owns their delivery sequence.