Merge pull request #88 from nold-ai/dev

Release dev to main: specfact-code-review 0.44.0

Author: djm81

CHANGELOG.md

@@ -5,6 +5,20 @@ All notable changes to this repository will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project follows SemVer for bundle versions.
 
+## [0.44.0] - 2026-03-17
+
+### Added
+
+- Add `--scope changed|full` and repeatable repo-relative `--path` filters to
+  `specfact code review run` for deterministic changed-only, full-repository,
+  and subtree-limited review selection.
+
+### Changed
+
+- Keep changed-only auto-discovery as the default, allow explicit test subtrees
+  to opt matching tests back into scope, and extend the review-run docs plus
+  cli-contract scenarios to cover the new targeting controls.
+
 ## [0.43.0] - 2026-03-16
 
 ### Added

docs/modules/code-review.md

@@ -17,6 +17,10 @@ Options:
 - `--score-only`: print only the integer `reward_delta`
 - `--fix`: apply Ruff autofixes and re-run the review before printing results
 - `--no-tests`: skip the targeted TDD gate
+- `--scope changed|full`: choose changed-only or full-repository auto-discovery
+  when positional files are omitted
+- `--path PATH`: repeatable repo-relative subtree filter for auto-discovered
+  review targets
 - `--include-tests/--exclude-tests`: include or exclude changed test files when
   review scope is auto-detected from `git diff`
 - `--include-noise/--suppress-noise`: include or suppress known low-signal
@@ -34,7 +38,27 @@ git ls-files --others --exclude-standard
 Only existing Python files from tracked or untracked workspace changes are
 reviewed. Test files under `tests/` are excluded by default for auto-detected
 review scope unless you pass `--include-tests` or answer yes in
-`--interactive` mode.
+`--interactive` mode. Explicit `--path tests/...` filters count as intentional
+targeting and keep matching tests in scope even with the default test
+exclusion.
+
+Use `--scope full` to review the governed repository file set instead of just
+current changes:
+
+```bash
+specfact code review run --scope full
+```
+
+Use repeatable `--path` filters to limit either scope to one package or a
+focused source-plus-test slice:
+
+```bash
+specfact code review run --scope full --path packages/specfact-code-review
+specfact code review run --scope changed --path packages/specfact-code-review --path tests/unit/specfact_code_review
+```
+
+Positional `FILES...` cannot be mixed with `--scope` or `--path`. Choose one
+targeting style per invocation.
 
 With default noise suppression, the review also hides known low-signal test
 findings such as:

openspec/changes/code-review-10-review-scope-modes/.openspec.yaml

@@ -0,0 +1,2 @@
+schema: spec-driven
+created: 2026-03-17

openspec/changes/code-review-10-review-scope-modes/CHANGE_VALIDATION.md

@@ -0,0 +1,54 @@
+## Verification Report: code-review-10-review-scope-modes
+
+### Summary
+
+| Dimension | Status |
+| --- | --- |
+| Completeness | 4/4 apply-ready artifacts present (`proposal`, `design`, `specs`, `tasks`) |
+| Correctness | Scope-mode and path-filter requirements align with upstream `code-review-10-review-scope-modes` |
+| Coherence | Proposal, design, and tasks consistently keep changed-only as the default auto-scope |
+
+### Validation Checks
+
+- Reviewed upstream `specfact-cli` proposal, design, and spec delta for `code-review-10-review-scope-modes`.
+- Reviewed local artifacts: `proposal.md`, `design.md`, `tasks.md`, and delta specs under `specs/`.
+- Checked current modules-repo review-run capabilities from `code-review-08-review-run-integration` to ensure this is a requirement change on an existing command, not a new command surface.
+
+### CRITICAL
+
+None.
+
+### WARNING
+
+- The derived change assumes the bundle command can identify a stable governed
+  repository file set for `--scope full`. If existing runtime logic only
+  supports changed-file discovery, implementation may need a new helper to keep
+  full-review selection deterministic across unit and e2e tests.
+  Recommendation: verify current target-resolution boundaries before coding and
+  centralize scope resolution rather than duplicating it across command tests
+  and runtime entry points.
+
+### SUGGESTION
+
+- Keep cli-val scenarios narrow and representative. One changed-only example,
+  one full-review example, one subtree-filter example, and one invalid
+  invocation are enough to prove the contract without making scenario YAML noisy.
+
+### Dependency Analysis
+
+- Modified capabilities: `review-run-command`, `review-cli-contracts`
+- Primary implementation touchpoint: `packages/specfact-code-review/src/specfact_code_review/run/`
+- Primary regression surface: `tests/unit/`, `tests/e2e/`, and `tests/cli-contracts/`
+- Upstream source dependency: `nold-ai/specfact-cli` change `code-review-10-review-scope-modes`
+
+### Final Assessment
+
+No critical issues found. The derived change is internally consistent, matches
+the upstream intent, and is ready for implementation after strict OpenSpec
+validation passes.
+
+### OpenSpec Validation
+
+- **Status**: Pass
+- **Command**: `openspec validate code-review-10-review-scope-modes --strict`
+- **Issues Found/Fixed**: 0

openspec/changes/code-review-10-review-scope-modes/TDD_EVIDENCE.md

@@ -0,0 +1,106 @@
+# TDD Evidence: code-review-10 review scope modes
+
+## Failing
+
+### 2026-03-17 `tests/unit/specfact_code_review/run/test_commands.py`
+
+Command:
+
+```bash
+HATCH_DATA_DIR=/tmp/hatch-data \
+HATCH_CACHE_DIR=/tmp/hatch-cache \
+VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata \
+hatch run pytest tests/unit/specfact_code_review/run/test_commands.py -q
+```
+
+Result: `5 failed, 14 passed`
+
+Key failures before implementation:
+
+- `test_run_command_supports_full_scope_and_path_filters`
+  - exit code was `2` because `--scope` was not supported yet
+- `test_run_command_supports_changed_scope_with_repeatable_path_filters`
+  - exit code was `2` because `--scope`/`--path` were not supported yet
+- `test_run_command_rejects_scope_mixed_with_positional_files`
+  - CLI returned the generic option error instead of the governed mixed-targeting message
+- `test_run_command_rejects_path_mixed_with_positional_files`
+  - CLI returned the generic option error instead of the governed mixed-targeting message
+- `test_run_command_fails_when_scope_and_paths_match_no_files`
+  - CLI returned the generic option error instead of an actionable empty-scope failure
+
+## Passing
+
+### 2026-03-17 Focused scope-mode tests
+
+Commands:
+
+```bash
+HATCH_DATA_DIR=/tmp/hatch-data \
+HATCH_CACHE_DIR=/tmp/hatch-cache \
+VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata \
+hatch run pytest tests/unit/specfact_code_review/run/test_commands.py tests/unit/specfact_code_review/run/test_runner.py -q
+
+HATCH_DATA_DIR=/tmp/hatch-data \
+HATCH_CACHE_DIR=/tmp/hatch-cache \
+VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata \
+hatch run pytest tests/e2e/specfact_code_review/test_review_run_e2e.py -q
+
+HATCH_DATA_DIR=/tmp/hatch-data \
+HATCH_CACHE_DIR=/tmp/hatch-cache \
+VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata \
+hatch run validate-cli-contracts
+```
+
+Results:
+
+- `tests/unit/specfact_code_review/run/test_commands.py` and `tests/unit/specfact_code_review/run/test_runner.py`: `36 passed`
+- `tests/e2e/specfact_code_review/test_review_run_e2e.py`: `2 passed`
+- CLI contract validation: `Validated 3 CLI contract scenario files.`
+
+### 2026-03-17 SpecFact dogfood review
+
+Command:
+
+```bash
+SPECFACT_ALLOW_UNSIGNED=1 \
+HATCH_DATA_DIR=/tmp/hatch-data \
+HATCH_CACHE_DIR=/tmp/hatch-cache \
+VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata \
+hatch run specfact code review run \
+  --scope changed \
+  --path packages/specfact-code-review \
+  --path tests/unit/specfact_code_review \
+  --json \
+  --out /tmp/code-review-10-report.json
+```
+
+Result:
+
+- verdict: `PASS`
+- score: `115`
+- findings: `0`
+
+### 2026-03-17 Worktree quality gates
+
+Commands completed successfully:
+
+```bash
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run format
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run type-check
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run lint
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run check-bundle-imports
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run smart-test
+HATCH_DATA_DIR=/tmp/hatch-data HATCH_CACHE_DIR=/tmp/hatch-cache VIRTUALENV_OVERRIDE_APP_DATA=/tmp/virtualenv-appdata hatch run contract-test
+```
+
+Results:
+
+- `type-check`: `0 errors, 0 warnings, 0 notes`
+- `lint`: `10.00/10`
+- `smart-test`: `383 passed`
+- `contract-test`: `384 passed`
+
+Pending:
+
+- `verify-modules-signature --require-signature --payload-from-filesystem --enforce-version-bump`
+  - blocked until the final `packages/specfact-code-review/module-package.yaml` changes in this worktree are re-signed

openspec/changes/code-review-10-review-scope-modes/design.md

@@ -0,0 +1,71 @@
+# Design: code-review scope modes for modules repo
+
+## Context
+
+`code-review-08-review-run-integration` gave the modules repo a working
+`specfact code review run` command, but its implicit changed-file discovery is
+too narrow for stepwise review debt cleanup. The upstream `specfact-cli`
+contract now defines two auto-discovery modes, `changed` and `full`, plus
+repo-relative path filters that can recursively narrow either mode.
+
+This repository owns the runtime command implementation, test fixtures, cli-val
+scenario YAML, and bundle docs. The design therefore focuses on deterministic
+file selection inside the bundle rather than on higher-level GitHub or
+automation wiring.
+
+## Goals / Non-Goals
+
+**Goals:**
+- Keep default no-argument behavior aligned with existing changed-file review
+- Add `--scope changed|full` and repeatable `--path` filters at the bundle
+  command boundary
+- Make subtree filtering recursive so users can review one package or test area
+  at a time
+- Cover the new behavior with unit, e2e, and cli-val fixtures
+
+**Non-Goals:**
+- Changing scoring, verdict mapping, or ledger integration
+- Adding globbing or exclusion syntax beyond repo-relative path prefixes
+- Reworking the internal runner ordering established in `code-review-08`
+
+## Decisions
+
+### Decision: Resolve scope before filtering by path
+
+The bundle should first compute the candidate file set from the selected scope,
+then filter that set by any `--path` prefixes. That keeps the behavior stable
+between tests and runtime, and makes it easy to explain to users.
+
+### Decision: Reject positional files mixed with scope controls
+
+Positional files already provide an explicit review target. Mixing them with
+`--scope` or `--path` would create ambiguous precedence rules, so the command
+should fail fast and tell the user to choose one targeting style.
+
+### Decision: Treat `--path` values as repo-relative recursive path segments
+
+Users asked for folder and subfolder limiting. Prefix-based matching is enough
+for that need, but matching must happen on normalized path boundaries rather
+than raw strings. That keeps `--path packages/specfact-code-review` scoped to
+that package and its descendants instead of accidentally including sibling
+paths such as `packages/specfact-code-review-old`, while still avoiding new
+pattern syntax in the CLI contract.
+
+### Decision: Explicit path filters may opt matched tests back in
+
+The existing changed-only workflow excludes tests by default unless users pass
+`--include-tests`. For scope-mode filtering, explicit `--path tests/...`
+selection should count as intentional targeting, so matching test files remain
+reviewable even when the global include-tests toggle stays at its default.
+
+## Risks / Trade-offs
+
+- [Risk] Full-review mode may surface far more findings than users expect.
+  Mitigation: keep it opt-in and document subtree filtering prominently.
+- [Risk] Scope/path selection may vary between unit tests and e2e execution if
+  repo root detection is inconsistent.
+  Mitigation: centralize target resolution in one helper and exercise it
+  through both direct tests and installed-command tests.
+- [Risk] cli-val scenarios could lag behind actual command behavior.
+  Mitigation: update review-run scenario YAML in the same change as the command
+  implementation and validate them together.

openspec/changes/code-review-10-review-scope-modes/proposal.md

@@ -0,0 +1,57 @@
+# Change: Reapply code-review-10 review scope modes in modules repo
+
+## Why
+
+The upstream `specfact-cli` change `code-review-10-review-scope-modes` defines
+the governed contract for explicit changed-only vs. full review modes plus
+repo-relative path filtering. The modules repository needs a matching active
+change because the bundle-owned `specfact code review run` implementation,
+runtime fixtures, and cli-val scenarios live here.
+
+Without this derived change, the upstream contract would exist only on paper:
+the installed `specfact-code-review` bundle would still lack the runtime
+behavior needed to review a full repo or a limited subtree in a controlled way.
+
+## What Changes
+
+- Extend the `specfact-code-review` bundle command to support `--scope changed`
+  and `--scope full`.
+- Add repeatable `--path` filtering for repo-relative files and subfolders in
+  auto-discovery mode.
+- Keep changed-only auto-discovery as the default behavior when users do not
+  pass positional files or an explicit scope.
+- Treat explicit `--path tests/...` filters as an intentional opt-in for those
+  matched test files without changing the default auto-scope exclusion of tests.
+- Add targeted tests, fixtures, and cli-val scenarios for changed-only reviews,
+  full reviews, filtered subtree reviews, and invalid scope combinations.
+- Update bundle docs, changelog, and signed package metadata to describe the
+  new review-scope controls.
+
+## Capabilities
+
+### New Capabilities
+
+None.
+
+### Modified Capabilities
+
+- `review-run-command`: add explicit review-scope selection and repo-relative
+  path filtering to the bundle command
+- `review-cli-contracts`: extend review-run scenario coverage for changed-only,
+  full-review, and subtree-limited invocations
+
+## Impact
+
+- Affects `packages/specfact-code-review/src/specfact_code_review/run/` command
+  parsing and target-file resolution
+- Expands runtime validation in `tests/unit/`, `tests/e2e/`, and
+  `tests/cli-contracts/`
+- Requires documentation, changelog, and module metadata updates for the
+  `specfact-code-review` bundle
+
+## Source Tracking
+
+<!-- source_repo: nold-ai/specfact-cli -->
+- **Source Change**: `code-review-10-review-scope-modes`
+- **Repository**: nold-ai/specfact-cli
+- **Last Synced Status**: re-applied in modules repo

openspec/changes/code-review-10-review-scope-modes/specs/review-cli-contracts/spec.md

@@ -0,0 +1,21 @@
+## MODIFIED Requirements
+
+### Requirement: cli-val scenarios exist for review command groups
+The modules repository SHALL define cli-val-compatible scenario YAML files for
+the `specfact code review run`, `ledger`, and `rules` command groups, including
+scope-mode and path-filter coverage for the review run command.
+
+#### Scenario: review-run scenarios cover success, scope selection, and error paths
+- **GIVEN** `tests/cli-contracts/specfact-code-review-run.scenarios.yaml`
+- **WHEN** it is validated
+- **THEN** it includes success coverage, changed-only and full-review scope examples, subtree-filtered examples, and an error or anti-pattern scenario
+
+#### Scenario: ledger scenarios cover update, status, and reset guardrails
+- **GIVEN** `tests/cli-contracts/specfact-code-review-ledger.scenarios.yaml`
+- **WHEN** it is validated
+- **THEN** it includes `update`, `status`, and reset guardrail coverage
+
+#### Scenario: rules scenarios cover init, show, and update
+- **GIVEN** `tests/cli-contracts/specfact-code-review-rules.scenarios.yaml`
+- **WHEN** it is validated
+- **THEN** it includes the supported rules subcommands