100%

Always-On Projection

Goal

Make read_matrix return a non-empty per-domain projection for ANY plan or bug whose ledger carries cycle entries, with no declared taxonomy — the thinnest end-to-end slice that proves the north-star for a plan. The render hook (_render_matrix) and read_matrix are already data-gated, so this section only has to make the projection non-empty.

Implementation Sketch

The single inert mechanism for activation is that a cycle entry carries no domain (the coordinate sentinel), so merge_coordinate/_coordinate_of (ledger_matrix.py:220,309) return {}/None and fold_matrix skips it. Default a missing domain to the synthetic constant _all, derive shape_id from the existing regex_coordinate_suggestion (keys off hypothesis/clause keywords), leave mechanism_class None. A degenerate single-cell-per-domain baseline is the honest floor; taxonomy + heal enrich it later. The default is a FALLBACK — entries with an inline domain or a sidecar coordinate are untouched (backward-compat with aims-burden-shaped data). TDD: the failing pin is written BEFORE the extractor change.

Spec References

  • .claude/rules/learning-ledger.md §15 (per-domain matrix read-view), §13.2 (read_ranked salience — NEVER recompute decay).
  • scripts/plan_corpus/ledger_matrix.py merge_coordinate, _coordinate_of, regex_coordinate_suggestion, fold_matrix, read_matrix.
  • Tests live in scripts/plan_corpus/tests/; pytest only (scripts scope — no compiler matrix/dual-exec).

Work Items

  • Write the failing pin first: a no-taxonomy fixture plan with >=3 ledger cycle entries asserts read_matrix returns a non-empty per-domain projection (TDD before code).
  • Default a missing domain to the synthetic constant _all inside merge_coordinate/_coordinate_of (ledger_matrix.py:220,309); derive shape_id via the existing regex_coordinate_suggestion; mechanism_class stays None.
  • Assert the inline default is a FALLBACK ONLY: entries carrying an inline domain or a sidecar coordinate are unchanged (backward-compat pin for aims-burden-tracking-shaped data).

Fresh intel (regenerated)

{ “schema_version”: 1, “target”: { “kind”: “plan-section”, “ref”: “ledger-matrix-replacement/s-55e12b57” }, “generated_at”: “2026-06-30T02:14:17.857188+00:00”, “graph_state”: { “head_sha”: “4ba25ed0”, “last_code_import_at”: “2026-06-30T02:14:15.397Z”, “embedding_stale”: false, “insights_stale”: false, “cpg_stale”: false }, “surfaces”: {}, “summary”: “intel-package[plan-section:ledger-matrix-replacement/s-55e12b57] agent-authored dossier”, “degraded”: false, “dossier”: { “objective_symbols”: [], “tiers”: {}, “agent_authored”: true, “constrained_symbols”: [ “assemble_corpus”, “read_ranked”, “validate_ledger_append_only”, “coordinate_heal_target”, “_plan_declares_matrix_taxonomy” ], “difficulty”: { “class”: “routine”, “research_online”: false, “research_mode”: “auto”, “signals”: [], “plan_dir”: “/home/eric/projects/ori_lang/plans/ledger-matrix-replacement” } } }

(full dossier: always-on-projection—s-55e12b57.intel.json)