fix(linker): restore raw similarity scoring

TODO.md

@@ -2,6 +2,11 @@
 
 ## Active tickets
 
+- [ ] [`ticket-031`](project/ticket-031/README.md) — restore the semantic
+  similarity score contribution dropped during the linker split. Current
+  state: `IN_PROGRESS / VALIDATION`; one production file changed, no threshold
+  or test expectation changes, and all seven targeted failures are removed.
+
 - [ ] [`ticket-019`](project/ticket-019/README.md) — publish the dependency-free
   Python SDK as the root PyPI distribution `todo2code` through `goal -a`, with
   one root `pyproject.toml` and SDK-only artifacts. Current state:

project/TICKETS.md

@@ -27,4 +27,5 @@ of `project/README.md`, which remains a generated technical-analysis artifact.
 | **ticket-019** | [`README.md`](./ticket-019/README.md) | [`preprompt.md`](./ticket-019/preprompt.md) | - |  [`ai-codex.md`](./ticket-019/ai-codex.md) |  [`ai-codex-logs.txt`](./ticket-019/ai-codex-logs.txt) | [`changelog.md`](./ticket-019/changelog.md) |
 | **ticket-020** | [`README.md`](./ticket-020/README.md) | [`preprompt.md`](./ticket-020/preprompt.md) | - |  [`ai-codex.md`](./ticket-020/ai-codex.md) |  [`ai-codex-logs.txt`](./ticket-020/ai-codex-logs.txt) | [`changelog.md`](./ticket-020/changelog.md) |
 | **ticket-022** | [`README.md`](./ticket-022/README.md) | [`preprompt.md`](./ticket-022/preprompt.md) | - |  [`ai-codex.md`](./ticket-022/ai-codex.md) |  [`ai-codex-logs.txt`](./ticket-022/ai-codex-logs.txt) | [`changelog.md`](./ticket-022/changelog.md) |
+| **ticket-031** | [`README.md`](./ticket-031/README.md) | [`preprompt.md`](./ticket-031/preprompt.md) | - |  [`ai-codex.md`](./ticket-031/ai-codex.md) |  [`ai-codex-logs.txt`](./ticket-031/ai-codex-logs.txt) | [`changelog.md`](./ticket-031/changelog.md) |
 <!-- AUTO:TICKET_INDEX:END -->

project/ticket-031/README.md

@@ -0,0 +1,67 @@
+# Ticket 031: Restore semantic similarity contribution after linker refactor
+
+- **ID**: ticket-031
+- **Owner**: unresolved:human
+- **Status**: IN_PROGRESS
+- **Workflow state**: VALIDATION
+- **Created**: 2026-08-04
+
+## Goal and scope
+
+Restore the semantic score contribution accidentally dropped while
+`scorePair` was split into focused helpers. The helper still computes and
+records object/text similarity, but the caller no longer adds its numeric
+value to the aggregate score. Exact path evidence therefore stops at `0.28`
+or `0.41` and valid cross-source relations fall below the `0.42` threshold.
+
+This is a bounded regression repair, not a change to the scoring model. The
+architecture remains: candidate selection in `linker-candidates.ts`, scoring
+in `linker.ts`, and relation direction in `linker-relations.ts`. Only
+`src/graph/linker.ts` may change. No threshold, fixture, gold expectation or
+generated analysis artifact may be edited. Expected implementation time is
+under 30 minutes.
+
+## Planned changed paths
+
+- `src/graph/linker.ts`: add the already computed object-similarity value to
+  the aggregate score exactly once.
+- `project/ticket-031/**`, `TODO.md`, `project/TICKETS.md`: intent and evidence.
+
+## Acceptance criteria
+
+- [x] AC-01: Scope is approved by a human owner.
+- [x] AC-02: `scorePair` adds object/text similarity to the aggregate score and
+      continues to expose the same value as `textScore`.
+- [x] AC-03: The shared-path plan/AST and configuration/document regression
+      tests pass without lowering the `0.42` relation threshold.
+- [x] AC-04: The versioned gold evaluation returns to its accepted precision,
+      recall, known-gap and v1-compatibility results.
+- [x] AC-05: The full suite reduces by the seven linker-related failures; no
+      unrelated test or fixture is weakened.
+- [x] AC-06: Clean build, focused tests, `git diff --check` and governance are
+      recorded with inherited blockers kept separate.
+
+## Participants
+
+- Human participant: unresolved; no user-* file was created by this script.
+- Agent participant: [ai-codex.md](ai-codex.md)
+
+## Approval boundary
+
+- Current state: `IN_PROGRESS / VALIDATION` after implementing and testing the
+  bounded repair.
+- Chat approval authorizes this interactive implementation only.
+- Trusted merge evidence still requires an independent protected review or
+  signed attestation bound to the final head SHA.
+
+## Validation evidence
+
+- `npm run build`: PASS.
+- Focused AST, linker-pairing and gold evaluation: 27 passed, 0 failed.
+- Full fresh suite: 338 tests, 333 passed, 4 failed and one skipped. The seven
+  linker failures are gone. The remaining failures are the three separately
+  diagnosed communication regressions plus the confidence-hierarchy test
+  repaired independently by ticket-030/PR #7.
+- `git diff --check`: PASS.
+- `make governance`: reports only the four inherited ticket-018/ticket-019
+  diagnostics; ticket-031 adds no scope or workstream finding.

project/ticket-031/ai-codex-logs.txt

@@ -0,0 +1,12 @@
+2026-08-04 planning
+- Fresh base `caf6551`: build PASS; full suite 338 total, 327 pass, 10 fail, 1 skip.
+- Seven failures trace to one omission: `scoreObjectSimilarity(...)` is computed but not added to aggregate `score` after the linker split.
+- No production source edited; state remains `WAIT_FOR_APPROVAL`.
+- Human response `kontynuuj` authorized the bounded ticket-031 implementation;
+  transitioned to `IN_PROGRESS / EDIT`.
+- `npm run build`: PASS.
+- Focused `ast`, `linker-pairing`, `gold-evaluation`: PASS (27/27).
+- Fresh full suite: 338 total, 333 pass, 4 fail, 1 skip; all seven linker
+  regressions removed. Remaining: three communication tests and ticket-030.
+- `git diff --check`: PASS.
+- `make governance`: four inherited ticket-018/ticket-019 errors, no ticket-031 finding.

project/ticket-031/ai-codex.md

@@ -0,0 +1,42 @@
+---
+participant-id: agent:codex
+participant: codex
+role: agent
+ticket: ticket-031
+---
+# Participant: codex (AI agent)
+
+## Understanding
+
+The refactor preserved `scoreObjectSimilarity(...)` and its evidence basis but
+assigned its return value only to `objectSimilarity`; the aggregate `score`
+never receives it. This single semantic omission explains seven fresh-suite
+failures: two direct linker cases, the AST integration case and four gold
+evaluation cases. Restoring the addition is safer than changing thresholds or
+test expectations because it recreates the pre-refactor algorithm.
+
+## Execution plan
+
+1. Wait for explicit human approval; do not edit production source beforehand.
+2. Add the computed similarity contribution exactly once in `scorePair`.
+3. Run build plus focused AST, linker-pairing and gold-evaluation tests.
+4. Run the full suite and verify that only the three separately diagnosed
+   communication failures remain.
+5. Run diff/scope checks and governance; publish only with truthful evidence.
+
+## Actual changes
+
+- The human explicitly approved ticket-031 by replying `kontynuuj`; the
+  interactive implementation may enter `EDIT`.
+- Restored the pre-refactor scoring contract in `src/graph/linker.ts`: raw
+  similarity remains available to relation classification, while its weighted
+  `0.48` contribution is added exactly once to the aggregate score.
+- Clean build and all 27 focused AST/linker/gold tests pass. The full suite no
+  longer contains any of the seven linker failures and stops only on three
+  communication regressions plus ticket-030's independently repaired test.
+
+## Blockers
+
+- Repository governance already has four unrelated ticket-018/ticket-019
+  blockers; ticket-031 must not rewrite them.
+- Trusted merge approval for the final head SHA is still required.

project/ticket-031/changelog.md

@@ -0,0 +1,16 @@
+# Ticket Changelog (ticket-031)
+
+## [0.1.0] - 2026-08-04
+
+- Initial governance scaffold created.
+- No human participant identity or content was generated.
+- Renumbered the unpublished scaffold from 023 to 031 because remote PR #3
+  already owns ticket-023.
+- Documented the pre-implementation regression evidence and bounded repair.
+
+## [0.2.0] - 2026-08-04
+
+- Restored raw object/text similarity as relation-classification evidence and
+  its weighted contribution to the aggregate linker score.
+- Removed seven fresh-suite regressions without changing thresholds, fixtures
+  or gold expectations; focused verification passes 27/27.

project/ticket-031/intent.json

@@ -0,0 +1,17 @@
+{
+  "schema": "new-project.intent/v2",
+  "ticket": "ticket-031",
+  "summary": "Restore semantic similarity contribution after linker refactor",
+  "workstream": "core-dsl",
+  "allowedPaths": [
+    "src/graph/linker.ts",
+    "project/ticket-031/**",
+    "TODO.md",
+    "project/TICKETS.md"
+  ],
+  "forbiddenPaths": ["project/ticket-*/user-*.md"],
+  "stacks": ["node"],
+  "dependsOn": [],
+  "conflictsWith": [],
+  "integrationTicket": null
+}

project/ticket-031/preprompt.md

@@ -0,0 +1,8 @@
+# Ticket preprompt
+
+- **Task ID**: ticket-031
+- **Task title**: Restore semantic similarity contribution after linker refactor
+- **Created**: 2026-08-04T12:07:14Z
+
+Keep executable implementation outside this governance/evidence directory.
+Read a human-owned user-*.md file only when one exists.

src/graph/linker.ts

@@ -151,6 +151,7 @@ function scorePair(
   score += scoreSharedPath(left, right, resolvableBasenames, basis);
   score += scoreSameAction(left, right, basis);
   const objectSimilarity = scoreObjectSimilarity(leftKeywords, rightKeywords, basis);
+  score += objectSimilarity * 0.48;
   score += scoreSharedTopics(left, right, leftKeywords, rightKeywords, basis);
   score += scoreSourceKindPenalty(left, right);
   return {
@@ -215,7 +216,7 @@ function scoreObjectSimilarity(
     : 0;
   if (objectSimilarity >= 0.2) {
     basis.push(`text_similarity:${objectSimilarity.toFixed(3)}`);
-    return objectSimilarity * 0.48;
+    return objectSimilarity;
   }
   return 0;
 }