Spec Kitty

└─ kitty-specs
   └─ CI Hardening and Lint Cleanup

Mission Run:

📚 Docs ↗

Implementation Plan: CI Hardening and Lint Cleanup

Branch: feat/080-ci-hardening-and-lint-cleanup | Date: 2026-04-09 Spec: spec.md | Research: research.md Mission type: software-dev

Summary

Bring the spec-kitty codebase to a clean ruff + mypy baseline, replace the monolithic CI test jobs with dependency-ordered per-module jobs (informed by a measured coverage baseline and tiered floor policy), and add path-based CI filtering so documentation-only commits do not trigger the full test suite.

The work is structured in three dependency-ordered execution batches:

Batch 1 (parallel): WP01–WP05 — lint and type fixes on main
Batch 2 (sequential after Batch 1): WP06 + WP07 + WP08 — baseline measurement, shift-left, and test marker cataloguing
Batch 3 (sequential after Batch 2): WP09 + WP10 — CI job split and path filtering

Technical Context

Language/Version: Python 3.11+ (runtime); Python 3.12 on CI runners Primary Dependencies: pytest, ruff, mypy, typer, rich, ruamel.yaml Storage: Filesystem only (YAML, JSONL, Markdown); no database Testing: pytest with markers: fast, git_repo, slow, integration, e2e Target Platform: GitHub Actions (ubuntu-24.04 runners); also Linux/macOS local dev Performance Goals: docs-only PRs ≤ 2 min CI; single-module code PRs ≤ 5 min CI Constraints: Branch protection required checks must not be broken at any point during migration (C-001) Scale/Scope: ~87,000 lines of Python source; ~13 module clusters; ~12 workflow files

Charter Check

Charter file: .kittify/charter/charter.md (present, v1.1.0)

Check	Status	Notes
mypy --strict passes	❌ (pre-mission) → ✅ (post-WP05)	WP01–WP05 fix all open violations
90%+ coverage for new code	✅ enforced by per-WP gate	New test code in WP08 must meet charter threshold
CLI < 2 seconds	not applicable	No runtime behavior changed
Terminology: `--mission` not `--feature`	not applicable	No CLI changes

No charter violations introduced by this mission.

Architectural Assessment (Architect Alphonso)

Profile: architect — analysis mode

Module Dependency Graph and CI DAG Ordering

The import-graph analysis (research.md R-01) confirms three execution tiers for the CI DAG:

Tier 0 (leaf nodes — run immediately):
  sync ─────────────────────────────────────────────────────┐
  merge ────────────────────────────────────────────────┐   │
  missions ─────────────────────────────────────────┐   │   │
  post_merge / release ─────────────────────────┐   │   │   │
                                                │   │   │   │
Tier 1 (depends on Tier 0):                     │   │   │   │
  status ◄───────────────────────────────────────── ────│───┘
                                                    │   │
Tier 2 (depends on Tier 1):                         │   │
  review ◄── status                                 │   │
  next   ◄── status                                 │   │
  lanes  ◄── status, merge ◄──────────────────────── ───┘
  dashboard ◄── status, sync ◄──────────────────────────┘
  upgrade ◄── status

Tier 3 (orchestrators — depend on Tier 2):
  cli ◄── {everything above}
  orchestrator_api ◄── cli, lanes, merge, status

Key architectural decision: Alphonso's original assumption ("next and review before merge") does not hold at the import level — merge has no imports from next or review. The actual structural ordering is sync → status → {review, next, lanes}. The CI DAG must encode this verified structure, not the intuitive but unverified assumption.

Coverage Floor Architecture

The three-tier floor model reflects the principle that coverage strictness should be proportional to the blast radius of a regression:

Essential (Tier A) — 75% minimum:
  status, lanes, kernel, sync
  Rationale: used by ≥3 modules; state-machine correctness is non-negotiable

Normal (Tier B) — 60% minimum:
  next, review, merge, cli, missions, upgrade
  Rationale: core workflow logic; bounded failure domain

Glue (Tier C) — 40% minimum:
  dashboard, release, orchestrator_api, post_merge, core-misc
  Rationale: thin adapters; integration tests cover the critical paths

Floors are set to max(tier_minimum, measured_baseline - 2%) — a "do not regress" gate, not a "achieve more coverage" mandate. Tightening is out of scope for this mission.

Shift-Left Assessment

The test pyramid is currently inverted in key modules:

marker-scoped CI job

tests/next/ has more git_repo tests than fast tests (5:4 ratio)
tests/missions/ has a 7:6 git_repo:fast ratio
tests/review/ and tests/lanes/ have zero marked tests — they are invisible to any

This creates two problems: 1. CI cost: git_repo tests spawn subprocess git init + git operations, costing ~30–90s per test. Tests that only need filesystem state should not pay this cost. 2. CI correctness: Unmarked tests in review and lanes are not counted in any marker- scoped coverage run — they contribute to line coverage but not to the job that reports it.

Architectural recommendation: FR-016 (marker cataloguing) is a prerequisite for FR-017 (shift-left) and must complete before WP09 (CI split) to ensure coverage accounting is correct from the first split run.

Risk Register (updated from original assessment)

Risk	Likelihood	Mitigation	Owner WP
Branch protection breaks on job rename	High	Add shim jobs before removing old required checks	WP10
Coverage floors set too high (miscalibration)	Medium	Measure-first + `-2%` buffer; conservative tier minimums	WP06
Dossier test fixes change test behavior	Medium	Inspect production signatures before editing; do not change test scenarios	WP03
Unmarked tests missed in coverage accounting	High	FR-016 cataloguing in WP07 before CI split in WP09	WP07
`dorny/paths-filter` output name collision	Low	Use namespaced output keys per job group	WP10
CI matrix explosion in `report`/`quality-gate`	Medium	Reusable workflow or job matrix for DAG enumeration	WP09

Doctrine References

Artifact	Type	Applied in
`DIRECTIVE_030` — Test and Typecheck Quality Gate	Directive	WP01–WP05 acceptance criteria
`DIRECTIVE_025` — Boy Scout Rule	Directive	WP02 (E402 cleanup scope)
`DIRECTIVE_034` — Test-First Development	Directive	WP08 (shift-left must not reduce coverage)
`DIRECTIVE_001` — Architectural Integrity Standard	Directive	WP09 (CI DAG must match architectural boundaries)
`test-pyramid-progression`	Tactic	WP09 job ordering; WP08 shift-left criteria
`testing-select-appropriate-level`	Tactic	WP08 classification of shift-left candidates
`test-boundaries-by-responsibility`	Tactic	WP09 job scoping (each job tests its own boundary)
`quality-gate-verification`	Tactic	WP06 floor calibration verification
`no-parallel-duplicate-test-runs`	Tactic	WP09 job split (no test must run in two jobs)

Project Structure

Documentation (this mission)

kitty-specs/080-ci-hardening-and-lint-cleanup/
├── spec.md              # Mission specification
├── plan.md              # This file
├── research.md          # Module graph, tier classification, shift-left findings
├── checklists/
│   └── requirements.md  # Spec quality checklist (all items pass)
└── tasks.md             # Generated by /spec-kitty.tasks (not yet created)

Source Code Touch Points

src/
├── charter/
│   ├── catalog.py             # WP01: ARG001 fix
│   └── resolver.py            # WP01: SIM108 fix
├── doctrine/missions/
│   └── glossary_hook.py       # WP01: B009 fix
├── kernel/
│   └── _safe_re.py            # WP01: SIM105 fix
└── specify_cli/
    ├── acceptance.py           # WP02: E402/UP035/F401 root-cause fix
    ├── dossier/tests/
    │   └── test_snapshot.py    # WP03: MissionDossier/ArtifactRef schema drift
    ├── post_merge/stale_assertions.py  # WP04: stale type: ignore (×4)
    ├── tracker/credentials.py  # WP04: stale type: ignore (×1)
    ├── merge/config.py         # WP04: stale type: ignore (×1)
    ├── migration/rebuild_state.py      # WP04: stale type: ignore (×1)
    ├── migration/backfill_identity.py  # WP04: stale type: ignore (×1)
    ├── policy/audit.py         # WP04: stale type: ignore (×1)
    ├── state_contract.py       # WP05: bare dict type-arg
    ├── acceptance_matrix.py    # WP05: bare dict type-arg (×2)
    ├── sync/config.py          # WP05: missing return type annotations
    ├── sync/body_transport.py  # WP05: missing types-requests stub
    ├── merge/conflict_resolver.py      # WP05: sort key type
    ├── migration/backfill_identity.py  # WP05: no-any-return
    ├── migration/backfill_ownership.py # WP05: bare dict type-arg
    ├── version_utils.py        # WP05: no-any-return
    ├── upgrade/feature_meta.py # WP05: no-any-return
    └── cli/commands/materialize.py     # WP05: incompatible assignment

tests/
├── next/          # WP08: shift-left candidates (5 git_repo tests)
├── missions/      # WP08: shift-left candidates (7 git_repo tests)
├── lanes/         # WP07: FR-016 marker cataloguing (0 markers on 12 test files)
└── review/        # WP07: FR-016 marker cataloguing (0 markers on 6 test files)

.github/workflows/
├── ci-quality.yml             # WP09 + WP10: job split + path filtering
├── orchestrator-boundary.yml  # WP10: FR-014 path filter
└── check-spec-kitty-events-alignment.yml  # WP10: FR-015 path filter

pyproject.toml                 # WP05: add types-requests to dev deps

Work Packages

Batch 1 — Lint and Type Fixes (parallel, independent)

WP01 — Auto-fix Ruff Violations

Scope: Fix all auto-fixable ruff violations in the four isolated-file locations. Relevant spec FRs: FR-001 Doctrine: DIRECTIVE_030, DIRECTIVE_025

Files to fix:

src/charter/catalog.py:245 — ARG001: remove or use the doctrine_root parameter
src/charter/resolver.py:120 — SIM108: collapse if/else to ternary
src/doctrine/missions/glossary_hook.py:134 — B009: replace getattr(module, "X") with module.X
src/kernel/_safe_re.py:185 — SIM105: replace try/except/pass with contextlib.suppress

Approach: Run ruff check --fix src/charter/catalog.py src/charter/resolver.py src/doctrine/missions/glossary_hook.py src/kernel/_safe_re.py for auto-fixable rules. Verify each change passes the test suite for that module.

Acceptance criteria:

ruff check src/charter/catalog.py src/charter/resolver.py src/doctrine/missions/glossary_hook.py src/kernel/_safe_re.py exits 0
All tests in the affected modules pass
DIRECTIVE_030 gate: mypy passes for touched files

WP02 — Fix acceptance.py Import Ordering

Scope: Resolve the E402 cascade in src/specify_cli/acceptance.py plus the UP035/F401 cluster caused by the same root issue. Relevant spec FRs: FR-001 Doctrine: DIRECTIVE_030, DIRECTIVE_025

Root cause: A logger = logging.getLogger(__name__) call appears at line 10, before all other imports. This produces 8 E402 violations on every subsequent import. The correct fix is to move the logger setup below the imports.

Files to fix:

extract_scalar, find_repo_root (unused); replace Dict/List/Set/Tuple/Iterable with their collections.abc or builtin equivalents (UP035)

src/specify_cli/acceptance.py: move logger init after imports; remove MutableMapping,

Acceptance criteria:

ruff check src/specify_cli/acceptance.py exits 0
mypy src/specify_cli/acceptance.py exits 0
All tests exercising acceptance.py pass (run pytest tests/ filtering by acceptance-touching test files)

WP03 — Fix Dossier Test Schema Drift

Scope: Fix src/specify_cli/dossier/tests/test_snapshot.py which calls MissionDossier and ArtifactRef with wrong keyword arguments. Relevant spec FRs: FR-002 Doctrine: DIRECTIVE_030, DIRECTIVE_034

Pre-condition (must verify first): Read the current production class signatures:

MissionDossier dataclass in src/specify_cli/dossier/ — determine current required fields
ArtifactRef dataclass — determine current required fields (wp_id, step_id, error_reason appear required per mypy)

Risk (Assumption A3): If the test scenarios test deprecated behavior rather than current behavior, do not fix them silently — escalate with a comment and flag for the reviewer.

Acceptance criteria:

mypy src/specify_cli/dossier/tests/test_snapshot.py exits 0 (zero call-arg and misc errors)
All tests in tests/specify_cli/ and the dossier test file pass
No production dataclass signatures are modified

WP04 — Remove Stale `# type: ignore` Comments

Scope: Remove 8 stale # type: ignore suppressions across 6 files. Relevant spec FRs: FR-002 Doctrine: DIRECTIVE_030

Files and lines:

src/specify_cli/post_merge/stale_assertions.py lines 317, 319, 322, 324
src/specify_cli/tracker/credentials.py line 15
src/specify_cli/merge/config.py line 57
src/specify_cli/migration/rebuild_state.py line 38
src/specify_cli/migration/backfill_identity.py line 36
src/specify_cli/policy/audit.py line 27

Verification process for each removal (C-005): 1. Remove the # type: ignore comment 2. Run mypy <file> — if an error reappears, the original suppression was valid; restore it and document the remaining error in a comment explaining why it persists

Acceptance criteria:

mypy for each of the 6 files reports zero unused-ignore errors
No new mypy errors introduced by the removals
All tests in affected modules pass

WP05 — Fix Remaining Mypy Violations

Scope: Fix bare generic types, no-any-return violations, missing return type annotations, incompatible assignments, and add types-requests dev dependency. Relevant spec FRs: FR-002, FR-003 Doctrine: DIRECTIVE_030

Grouped by fix type:

Bare generics (dict → dict[str, ...], etc.):

src/specify_cli/state_contract.py:80
src/specify_cli/acceptance_matrix.py:75,88
src/doctrine/missions/repository.py:56,74
src/specify_cli/migration/backfill_ownership.py:86,151

Missing return types:

src/specify_cli/sync/config.py:15,39 — add -> None

no-any-return (functions declared to return a specific type but return Any):

src/specify_cli/version_utils.py:28
src/specify_cli/upgrade/feature_meta.py:165
src/specify_cli/policy/audit.py:27 (also has a stale ignore — coordinate with WP04 if same WP)
src/doctrine/missions/repository.py:31,36,66,71,76
src/specify_cli/migration/backfill_identity.py:36 (coordinate with WP04)

Type incompatibilities:

src/specify_cli/merge/conflict_resolver.py:172 — sort key callable type
src/specify_cli/cli/commands/materialize.py:123 — Any | None assigned to str
src/specify_cli/tracker/credentials.py:17 — None assigned to Module type

Missing stubs (C-006: dev deps only):

Add types-requests to pyproject.toml under [project.optional-dependencies] lint or test group (not runtime)

Acceptance criteria:

mypy --strict src/ reports zero errors for all files listed in FR-002
All existing tests pass
types-requests is in dev/lint dependencies only

Batch 2 — Baseline Measurement and Test Classification (sequential after Batch 1)

Batch 2 starts only after all of WP01–WP05 are merged to the feature branch.

WP06 — Measure Per-Module Coverage Baseline

Scope: Run the full test suite (all markers) per module cluster and record actual coverage percentages. Produce the floor table for WP09. Relevant spec FRs: FR-006 Doctrine: quality-gate-verification, DIRECTIVE_030

Process: For each module cluster listed in the Tier table (research.md R-02):

pytest tests/<module>/ \
  --cov=src/specify_cli/<module> \
  -m 'fast or git_repo or slow or integration or e2e' \
  --cov-report=term-missing -q

Record results in kitty-specs/080-ci-hardening-and-lint-cleanup/coverage-baseline.md.

Floor formula per module:

floor = max(tier_minimum, measured_coverage - 2)

Where tier minimums are:

Essential (status, lanes, kernel, sync): 75%
Normal (next, review, merge, cli, missions, upgrade): 60%
Glue (dashboard, release, orchestrator_api, post_merge, core-misc): 40%

Acceptance criteria:

All module clusters have a recorded measured coverage percentage
coverage-baseline.md is committed to the feature branch
No test failures during baseline run (if failures occur, halt and report — do not proceed to WP07/WP08/WP09 with failing tests)

WP07 — Test Marker Cataloguing (FR-016)

Scope: Audit all test files in tests/lanes/ and tests/review/ (zero current markers) and any other module directories where a significant number of test functions are unmarked. Add appropriate pytest markers. Relevant spec FRs: FR-016 Doctrine: testing-select-appropriate-level, test-pyramid-progression

Classification criteria:

fast: no subprocess, no git, no network, no disk writes outside tmp_path — pure logic
git_repo: requires real git repository (subprocess git init or worktree)
slow: takes > 30 seconds
integration: requires multiple real components interacting
e2e: exercises the full CLI workflow end-to-end

Scope: 1. tests/lanes/ (12 test files, 0 markers) — classify all test functions 2. tests/review/ (6 test files, 0 markers) — classify all test functions 3. tests/merge/ (7 test files, 2 fast, 1 git_repo) — classify unmarked functions 4. tests/cli/ (3 test files, 0 markers) — classify all

Acceptance criteria:

Zero test functions in the catalogued modules lack at least one pytest marker
All tests still pass after marker addition (markers are additive, not behavioral)
pytest tests/lanes/ tests/review/ tests/merge/ tests/cli/ -m fast runs without error

WP08 — Shift-Left Test Migration (FR-017)

Scope: Convert git_repo-marked tests in tests/next/ and tests/missions/ that do not require real git operations to fast-marked tests. Relevant spec FRs: FR-017 Doctrine: testing-select-appropriate-level, test-pyramid-progression, DIRECTIVE_034

Identification criteria for shift-left eligibility: A git_repo test is a shift-left candidate if: 1. It does not call subprocess.run/subprocess.check_call/check_output with git commands 2. It does not use git worktree, git init, git commit, or any git operation that requires a real repo 3. Its assertions are on data structures, rendered content, or file contents — not on git state

Process: 1. For each git_repo test in tests/next/ and tests/missions/:

Check if git operations can be replaced with tmp_path + fixture data
If yes: refactor to use tmp_path, replace @pytest.mark.git_repo with @pytest.mark.fast
If no: document why it must remain git_repo in a comment on the test

2. Verify the converted tests still assert the same outcomes

Acceptance criteria (DIRECTIVE_034 — test-first):

No test coverage decreases after shift-left (run full suite before and after; compare coverage)
Converted tests run in the fast marker tier (confirmed by pytest -m fast tests/next/ tests/missions/)
Each converted test still passes and asserts the same invariant as before
Any test that cannot be shifted has a comment explaining why

Batch 3 — CI Structure (sequential after Batch 2)

Batch 3 starts only after WP06 (coverage-baseline.md produced) and WP07 (all tests marked).

WP09 — Per-Module CI Job Split (FR-004 to FR-009)

Scope: Replace fast-tests-core and integration-tests-core in .github/workflows/ci-quality.yml with per-module job pairs, using the dependency DAG from R-01 and coverage floors from WP06. Relevant spec FRs: FR-004, FR-005, FR-006, FR-007, FR-008, FR-009 Doctrine: DIRECTIVE_001, test-boundaries-by-responsibility, no-parallel-duplicate-test-runs

Job naming convention: fast-tests-<module> and integration-tests-<module>

DAG encoding in GitHub Actions:

# Tier 0 jobs — no needs:
fast-tests-sync:
fast-tests-merge:
fast-tests-missions:
fast-tests-post_merge:
fast-tests-release:

# Tier 1 — needs Tier 0:
fast-tests-status:
  needs: [fast-tests-sync]

# Tier 2 — needs Tier 1:
fast-tests-review:
  needs: [fast-tests-status]
fast-tests-next:
  needs: [fast-tests-status]
fast-tests-lanes:
  needs: [fast-tests-merge, fast-tests-status]
fast-tests-dashboard:
  needs: [fast-tests-status]
fast-tests-upgrade:
  needs: [fast-tests-status]

# Tier 3 — needs Tier 2:
fast-tests-cli:
  needs: [fast-tests-lanes, fast-tests-next, fast-tests-review, fast-tests-dashboard, fast-tests-upgrade]
fast-tests-orchestrator_api:
  needs: [fast-tests-cli, fast-tests-lanes, fast-tests-merge, fast-tests-status]

# Residual catch-all for unlisted modules:
fast-tests-core-misc:
  # tests/core/, tests/policy/, tests/schemas/, tests/validators/, etc.

Coverage floor application (from WP06 coverage-baseline.md):

- name: Check coverage floor
  run: |
    coverage report --fail-under=<floor_from_baseline>

report and quality-gate DAG update (FR-009): List all new per-module job names explicitly in the needs: array of both jobs.

Acceptance criteria:

fast-tests-core and integration-tests-core jobs no longer exist
All new per-module jobs run and pass on main
quality-gate job passes on main
DIRECTIVE_030 gate: full suite passes

WP10 — Path-Based CI Filtering + Branch Protection Shims (FR-010 to FR-015)

Scope: Add path-based workflow and job filters; add skip-pass shim jobs; update orchestrator-boundary.yml and check-spec-kitty-events-alignment.yml. Relevant spec FRs: FR-010, FR-011, FR-012, FR-013, FR-014, FR-015 Doctrine: DIRECTIVE_001

Pre-condition (C-001): Confirm with the repository owner which jobs are currently listed as required checks in branch protection before making any changes.

Step 1 — Add workflow-level paths: trigger to ci-quality.yml:

on:
  pull_request:
    branches: [main, develop, 2.x]
    paths:
      - 'src/**'
      - 'tests/**'
      - 'pyproject.toml'
      - '.github/workflows/ci-quality.yml'
  push:
    branches: [main, develop, 2.x]
    paths:
      - 'src/**'
      - 'tests/**'
      - 'pyproject.toml'
      - '.github/workflows/ci-quality.yml'

Step 2 — Add changes detection job using dorny/paths-filter:

changes:
  runs-on: ubuntu-latest
  outputs:
    sync: ${{ steps.filter.outputs.sync }}
    merge: ${{ steps.filter.outputs.merge }}
    # ... all modules
  steps:
    - uses: dorny/paths-filter@v3
      id: filter
      with:
        filters: |
          sync:
            - 'src/specify_cli/sync/**'
            - 'tests/sync/**'
          merge:
            - 'src/specify_cli/merge/**'
            - 'tests/merge/**'
          # ... etc.

Step 3 — Add if: conditions to each per-module job:

fast-tests-sync:
  if: needs.changes.outputs.sync == 'true'
  needs: [changes]

Step 4 — Add skip-pass shim jobs (FR-013): For each per-module job that is a required check, add:

fast-tests-<module>-shim:
  runs-on: ubuntu-latest
  needs: [changes]
  if: needs.changes.outputs.<module> != 'true'
  steps:
    - run: echo "Skipped — no <module> changes"

Step 5 — Update orchestrator-boundary.yml (FR-014): Add path filter: src/specify_cli/orchestrator_api/, src/specify_cli/core/

Step 6 — Update check-spec-kitty-events-alignment.yml (FR-015): Add path filter: src/specify_cli/sync/**, pyproject.toml

Acceptance criteria:

A PR touching only docs/** does not trigger Python test jobs
A PR touching only src/specify_cli/dashboard/ does not trigger fast-tests-merge or fast-tests-sync
All required checks remain satisfied (shim jobs or skip-pass pattern active)
orchestrator-boundary.yml does not run on a docs-only PR
check-spec-kitty-events-alignment.yml does not run on a docs-only PR

Execution Batches and Lane Strategy

Lane A (parallel):  WP01 → merge
Lane B (parallel):  WP02 → merge
Lane C (parallel):  WP03 → merge
Lane D (parallel):  WP04 → merge
Lane E (parallel):  WP05 → merge

  ↓ (all of Batch 1 merged)

Lane F (sequential): WP06 → WP07 → WP08 → merge

  ↓ (Batch 2 merged, coverage-baseline.md committed)

Lane G (sequential): WP09 → WP10 → merge

Rationale for sequential Batch 2: WP06 (measurement) must precede WP07 (marker cataloguing) because marker addition changes which tests run in marker-scoped coverage commands. WP08 (shift-left) must follow WP07 because some tests being shifted may gain or lose the fast marker during cataloguing.

Rationale for sequential Batch 3: WP09 (job split) must precede WP10 (path filtering) because the path filter's per-module if: conditions reference job names that WP09 creates.

Branch Contract (repeated per skill requirement)

Current branch at plan start: feat/080-ci-hardening-and-lint-cleanup
Planning/base branch: feat/080-ci-hardening-and-lint-cleanup
Final merge target for completed changes: main
branch_matches_target: true (we are on the feature branch throughout)

Next step: /spec-kitty.tasks — generates work package files and finalizes lane assignments.

Spec Kitty

Implementation Plan: CI Hardening and Lint Cleanup

Summary

Technical Context

Charter Check

Architectural Assessment (Architect Alphonso)

Module Dependency Graph and CI DAG Ordering

Coverage Floor Architecture

Shift-Left Assessment

Risk Register (updated from original assessment)

Doctrine References

Project Structure

Documentation (this mission)

Source Code Touch Points

Work Packages

Batch 1 — Lint and Type Fixes (parallel, independent)

WP01 — Auto-fix Ruff Violations

WP02 — Fix acceptance.py Import Ordering

WP03 — Fix Dossier Test Schema Drift

WP04 — Remove Stale # type: ignore Comments

WP05 — Fix Remaining Mypy Violations

Batch 2 — Baseline Measurement and Test Classification (sequential after Batch 1)

WP06 — Measure Per-Module Coverage Baseline

WP07 — Test Marker Cataloguing (FR-016)

WP08 — Shift-Left Test Migration (FR-017)

Batch 3 — CI Structure (sequential after Batch 2)

WP09 — Per-Module CI Job Split (FR-004 to FR-009)

WP10 — Path-Based CI Filtering + Branch Protection Shims (FR-010 to FR-015)

Execution Batches and Lane Strategy

Branch Contract (repeated per skill requirement)

WP04 — Remove Stale `# type: ignore` Comments