Economics-Accounting Eval-Ready Implementation Plan

For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (- [ ]) syntax for tracking.

Goal: Move the economics-accounting bridge from a deferred candidate shell to eval_ready with bridge-owned fixture coverage, manifest-backed signals, method-only borrowing, mixed-adjacent guards, and near-miss protection.

Architecture: Keep economics-accounting inactive in default runtime until a later promotion review. Use evaluation_subjects=["economics-accounting"] under the eval-ready gate to measure bridge behavior. Avoid built-in economics trigger phrases such as DID, causal identification, policy shock, and identification strategy, and avoid broad accounting subject phrases such as financial reporting, earnings quality, and disclosure quality in clear-positive bridge fixtures so the bridge tests measure the composite manifest rather than pre-existing economics/accounting subjects.

Tech Stack: Python 3, unittest, JSON router fixtures, YAML runtime subject manifests, Qiongli subject router evaluation tooling.

---

File Map

• Modify: tests/test_subject_router_eval.py
  • Add economics-accounting fixture inventory coverage.
  • Add a real fixture-pack eval-ready gate test.
  • Add a runtime-enabled guard proving checked-in eval_ready is not default activation.
  • Remove economics-accounting from the deferred-shell candidate loop.
• Modify: tests/test_subject_contracts.py
  • Classify economics-accounting as eval_ready instead of a deferred candidate.
  • Add manifest structure checks for bridge signal groups, method lenses, blank subject skill, evaluation pack, and metrics.
• Modify: content/subjects/economics-accounting/runtime-subject.yaml
  • Set activation_status: eval_ready.
  • Keep subject_skill: "" because there is no subject-owned skill yet and eval-ready allows blank optional subject resources.
  • Add method, data/outcome, venue, and theory/construct signal groups.
  • Add method lenses and evaluation_pack: tests/fixtures/subject_router_eval/economics-accounting.
• Create: tests/fixtures/subject_router_eval/economics-accounting/*.json
  • Add clear positive, method-only borrow, confirmed-subject, mixed-adjacent, and near-miss fixtures.
• Modify: docs/superpowers/roadmaps/2026-07-01-adaptive-subject-runtime-roadmap.md
  • Record economics-accounting as eval-ready.
  • Leave runtime activation as a separate follow-up.

Execution Notes

• Do not promote economics-accounting to runtime_enabled in this slice.
• Do not change accounting, economics, finance, business, political-economy, or geoeconomics routing behavior.
• Do not modify packages/python-qiongli/src/qiongli/bridges/subject_refinement.py unless the measured RED/GREEN path proves the existing runtime-subject contract cannot express the bridge.
• Keep venue entries context-only and avoid venue-only method lenses so course/syllabus references to accounting venues do not borrow bridge lenses.
• Keep commits grouped by behavior and docs:
  • feat(subjects): mark economics-accounting eval ready
  • docs(roadmap): record economics-accounting eval readiness

Task 1: Add Economics-Accounting Gate Tests

Files:

• Modify: tests/test_subject_router_eval.py

• Modify: tests/test_subject_contracts.py

• [ ] Step 1: Add fixture inventory expectations

Add this required fixture ID block next to the other subject inventory blocks:

        required_economics_accounting_ids = {
            "economics_accounting_clear_identification_measurement_bridge",
            "economics_accounting_method_only_alignment_borrow",
            "economics_accounting_confirmed_bridge_positioning",
            "economics_accounting_mixed_accounting_economics_bridge",
            "economics_accounting_near_miss_course_comparison",
            "economics_accounting_near_miss_internal_reporting_workflow",
        }
        self.assertTrue(required_economics_accounting_ids.issubset(set(ids)))
        economics_accounting_tags = {
            tag
            for case_id in required_economics_accounting_ids
            for tag in list(cases_by_id[case_id].tags or [])
        }
        self.assertTrue(
            {
                "clear_positive",
                "method_only_borrow",
                "mixed_subject",
                "near_miss",
                "confirmed_subject",
            }.issubset(economics_accounting_tags)
        )

• [ ] Step 2: Add eval-ready gate test

Add:

    def test_economics_accounting_eval_ready_gate_passes_real_fixture_pack(
        self,
    ) -> None:
        cases = load_eval_cases(FIXTURE_DIR)

        report = subject_gate_report("economics-accounting", cases, gate="eval-ready")

        self.assertEqual(report["subject"], "economics-accounting")
        self.assertEqual(report["activation_status"], "eval_ready")
        self.assertTrue(report["eligible_for_eval_ready"])
        self.assertFalse(report["eligible_for_runtime_enabled"])
        self.assertEqual(report["blocking_failures"], [])
        self.assertEqual(report["metrics"]["decision_accuracy"], 1.0)
        self.assertEqual(report["metrics"]["primary_subject_accuracy"], 1.0)
        self.assertEqual(report["metrics"]["suggest_subject_precision"], 1.0)
        self.assertEqual(report["metrics"]["forbidden_subject_accuracy"], 1.0)
        self.assertEqual(report["metrics"]["method_lens_accuracy"], 1.0)
        self.assertEqual(report["metrics"]["all_case_checks_passed"], 1.0)
        self.assertEqual(report["metrics"]["near_miss_false_positives"], 0)

• [ ] Step 3: Add runtime-enabled guard

Add:

    def test_economics_accounting_runtime_enabled_gate_blocks_eval_ready_manifest(
        self,
    ) -> None:
        cases = load_eval_cases(FIXTURE_DIR)

        report = subject_gate_report(
            "economics-accounting",
            cases,
            gate="runtime-enabled",
        )

        self.assertEqual(report["activation_status"], "eval_ready")
        self.assertFalse(report["eligible_for_eval_ready"])
        self.assertFalse(report["eligible_for_runtime_enabled"])
        self.assertIn("activation_status is eval_ready", report["blocking_failures"])

• [ ] Step 4: Update deferred candidate loop

Change the deferred subject tuple in test_candidate_subject_eval_ready_gate_reports_deferred_shell_reasons to:

        deferred_subjects = ()

• [ ] Step 5: Update contract classification

In tests/test_subject_contracts.py, classify economics-accounting as eval_ready and remove it from the deferred candidate shell set. Add a manifest test checking:

contract.activation_status == "eval_ready"
contract.evaluation_pack == "tests/fixtures/subject_router_eval/economics-accounting"
contract.subject_skill == ""

The manifest test must also assert all four signal dimensions are populated, method lenses include identification-measurement-alignment, fiscal-window-alignment, and economics-accounting-positioning, and gate metrics remain 0.95, 0.95, and 0.

• [ ] Step 6: Run RED

Run:

.venv/bin/python -m unittest tests.test_subject_router_eval.SubjectRouterEvalTests.test_economics_accounting_eval_ready_gate_passes_real_fixture_pack -q

Expected before implementation:

FAIL: activation_status is candidate

Task 2: Add Economics-Accounting Runtime Manifest

Files:

• Modify: content/subjects/economics-accounting/runtime-subject.yaml

• [ ] Step 1: Promote only to eval-ready

Set:

activation_status: eval_ready
subject_skill: ""
evaluation_pack: tests/fixtures/subject_router_eval/economics-accounting

• [ ] Step 2: Add method-only signals and lenses

Use method-only signals for:

• identification-measurement-alignment
• fiscal-window-alignment

Method lenses:

• identification-measurement-alignment -> content/subjects/economics-accounting/overlays/skills/stats-engine.md

• fiscal-window-alignment -> content/subjects/economics-accounting/overlays/skills/stats-engine.md

• economics-accounting-positioning -> content/subjects/economics-accounting/overlays/skills/manuscript-architect.md

• [ ] Step 3: Add subject-level signals

Subject-level signals must include:

• data/outcome: accounting disclosure rule, disclosure institution, fiscal-year timing, capital-market outcome windows, archival firm-year panel.
• theory/construct: economics and accounting contribution, accounting measurement and economic variation, cross-disciplinary contribution, composite reviewer risk.

Venue signals should include Journal of Accounting Research, The Accounting Review, and Review of Accounting Studies, but should not attach method lenses.

Task 3: Add Economics-Accounting Fixtures

Files:

• Create: tests/fixtures/subject_router_eval/economics-accounting/clear_identification_measurement_bridge.json

• Create: tests/fixtures/subject_router_eval/economics-accounting/method_only_alignment_borrow.json

• Create: tests/fixtures/subject_router_eval/economics-accounting/confirmed_bridge_positioning.json

• Create: tests/fixtures/subject_router_eval/economics-accounting/mixed_accounting_economics_bridge.json

• Create: tests/fixtures/subject_router_eval/economics-accounting/near_miss_course_comparison.json

• Create: tests/fixtures/subject_router_eval/economics-accounting/near_miss_internal_reporting_workflow.json

• [ ] Step 1: Add clear positive fixture

Use Journal of Accounting Research, economics-and-accounting contribution, identification-measurement alignment, estimand measurement proxy map, identifying variation, accounting disclosure rule, disclosure institution, fiscal-year timing, and capital-market outcome windows. Under eval-ready, expect:

{
  "decision": "recommend",
  "primary_subject": "economics-accounting",
  "suggest_subjects": ["economics-accounting"],
  "forbidden_subjects": [],
  "method_lenses": [
    "identification-measurement-alignment",
    "economics-accounting-positioning"
  ]
}

Default expected should suppress subject suggestion and only borrow bridge lenses until runtime activation.

• [ ] Step 2: Add method-only borrow fixture

Use only identification-measurement alignment and estimand measurement proxy map wording without subject-level data, venue, or theory signals. Expect primary auto, no economics-accounting suggestion, and method lens identification-measurement-alignment.

• [ ] Step 3: Add confirmed-subject fixture

Use active_subject: "economics-accounting" and subject_mode: "confirmed". Expect primary economics-accounting and no inferred suggestion requirement.

• [ ] Step 4: Add mixed-adjacent fixture

Use an accounting venue and bridge-specific subject evidence, but avoid built-in economics method trigger phrases. Under eval-ready, expect economics-accounting primary. Allow accounting and economics only as neighbors if they appear.

• [ ] Step 5: Add near-miss fixtures

Course comparison and internal reporting workflow near misses must remain core_only, with forbidden_subjects: ["economics-accounting"].

Task 4: Verify And Update Roadmap

Files:

• Modify: docs/superpowers/roadmaps/2026-07-01-adaptive-subject-runtime-roadmap.md

• [ ] Step 1: Run GREEN checks

Run:

.venv/bin/python -m unittest tests.test_subject_router_eval.SubjectRouterEvalTests.test_economics_accounting_eval_ready_gate_passes_real_fixture_pack -q
.venv/bin/python tooling/scripts/evaluate_subject_router.py --subject economics-accounting --gate eval-ready --json

Expected:

eligible_for_eval_ready: true
blocking_failures: []
near_miss_false_positives: 0

• [ ] Step 2: Run regressions

Run:

.venv/bin/python -m unittest tests.test_subject_router_eval tests.test_subject_contracts -q
.venv/bin/python tooling/scripts/evaluate_subject_router.py --json
git diff --check

Expected: tests and full fixture evaluation exit 0; whitespace check exits 0.

• [ ] Step 3: Commit by content

Run:

git add content/subjects/economics-accounting/runtime-subject.yaml tests/test_subject_router_eval.py tests/test_subject_contracts.py tests/fixtures/subject_router_eval/economics-accounting
git commit -m "feat(subjects): mark economics-accounting eval ready"
git add docs/superpowers/plans/2026-07-06-economics-accounting-eval-ready.md docs/superpowers/roadmaps/2026-07-01-adaptive-subject-runtime-roadmap.md
git commit -m "docs(roadmap): record economics-accounting eval readiness"