feat(pre-push-safety-net): connect change context to ranked questions

.kiro/specs/pre-push-safety-net/implementation-group-3.md

@@ -90,3 +90,39 @@ This manually declared entry-region smoke does not validate mapper discovery,
 PMPro/NodeGoat replay, capability admission or hook latency.
 
 Nine of 27 tasks are now complete; tasks 3.3–3.4 remain for this group.
+
+## Task 3.3 evidence
+
+Independent review APPROVED after one required repair; verification status VERIFIED,
+claim type TASK. The fresh reviewer suite passed 1,128 tests, nine skipped, in
+105.95 seconds. Global Ruff, formatting (224 files), mypy (108 source files) and
+diff checks passed. The context suite contains 23 focused cases.
+
+The generic driver accepts `change_context` and retains its enriched form. Existing
+taint evidence optionally projects full first-party method paths/ranges for entry,
+callee, same-file field, known hook and sanitizer context. Decoded snapshot changes
+and lexical anchors attach to exact questions without changing ranking weights or
+introducing another scope selector. Missing base, locations, empty projections,
+dynamic/external/depth dependencies and unsupported configuration/deleted-dependency
+projections remain explicit gaps. Missing context prevents tier-zero pruning and
+complete-negative outcomes; raw answers remain available.
+
+Review found that broad source/sink/sanitizer text matching could mistakenly exempt
+an unknown consumer such as `unknown(req.body)` from missing-context reporting.
+The production-query regression reproduced exit 1 with that question incorrectly
+pruned as tier zero. The task-local repair restricts the exemption to directly
+invoked fact-modeled operations. Existing detector matchers are unchanged. After
+repair, consumers of a source, nested sink and sanitizer name all remain unresolved;
+a direct modeled sink completes, and the direct sanitizer retains tier zero without
+a false context gap.
+
+`benchmarks/measurements/2026-09-13-change-context-smoke.json` retains the final image
+and matching backend/driver/regions/snapshot/query hashes, plus both runtime checks.
+Real Git histories remove a caller guard while keeping helper code unchanged. PHP
+reads 124 source bytes and reaches `helper.php:2:helper` in 63.48 seconds; JavaScript
+reads 133 bytes and reaches `helper.js:1:helper` through an ES-module import in
+39.63 seconds. Both preserve the dirty checkout/index and remove snapshot scratch.
+These are controlled change-to-question/witness checks, not proof that the removed
+line enforced security, that the defect is new, or that a hook meets its latency bar.
+
+Ten of 27 tasks are now complete; task 3.4 remains for this group.

.kiro/specs/pre-push-safety-net/spec.json

@@ -1,7 +1,7 @@
 {
   "feature_name": "pre-push-safety-net",
   "created_at": "2026-09-12T15:35:00Z",
-  "updated_at": "2026-09-13T08:16:14.274231+00:00",
+  "updated_at": "2026-09-13T08:41:28.781759+00:00",
   "language": "en",
   "phase": "implementation",
   "approvals": {
@@ -71,9 +71,10 @@
       "2.2",
       "2.3",
       "3.1",
-      "3.2"
+      "3.2",
+      "3.3"
     ],
-    "remaining_tasks": 18,
+    "remaining_tasks": 17,
     "verification_record": "implementation-group-3.md",
     "verification_records": [
       "implementation-group-1.md",

.kiro/specs/pre-push-safety-net/tasks.md

@@ -1,7 +1,7 @@
 # Implementation Plan: pre-push-safety-net
 
 Status: task plan approved by the maintainer's "approve." on 2026-09-12. Groups 1 and 2 (tasks 1.1–2.3)
-are complete after independent review and fresh verification. Group 3 is in progress (3.1–3.2 verified); groups 4–8 remain pending;
+are complete after independent review and fresh verification. Group 3 is in progress (3.1–3.3 verified); groups 4–8 remain pending;
 the hook and capability admission are not yet implemented. See `implementation-group-1.md` and `implementation-group-2.md`.
 
 ## Execution contract
@@ -95,13 +95,14 @@ the hook and capability admission are not yet implemented. See `implementation-g
   - _Boundary: Ranker Scope Contract, generic driver integration_
   - _Requirements: 2.1, 2.2, 6.4, 7.3, 8.1, 8.2_
   - _Verified 2026-09-13:_ independent review APPROVED; 1,105 passed, nine skipped; global Ruff/format/mypy passed. Twenty-four independent order/budget/missing-family combinations retained all question identities. Packaged PHP/JavaScript smoke selected the same controlled risky operation at tier 4/score 2.5, with exact execution/outcome IDs and reconciled family coverage. Missing evidence, partial graphs and interrupted arbitration remain explicit; CLI consumes authoritative scope. Evidence: `benchmarks/measurements/2026-09-13-ranker-scope-smoke.json`. No new ranking weights or hook admission claim.
-- [ ] 3.3 Connect change context to supported first-party relationships
+- [x] 3.3 Connect change context to supported first-party relationships
   - Use existing entry/callee, field, hook and guard relationships to supply affected context to ranked questions, including unchanged sinks after source/guard changes.
   - Preserve current abstract semantics; absent dynamic or external relationships remain unresolved rather than guessed safe. A required new feature or fact is reported to its owning spec.
   - Done when controlled cross-file and removed-guard cases reach the relevant question/witness, while an unresolved boundary cannot produce a complete negative.
   - _Boundary: Ranker Scope Contract, Snapshot Adapter integration_
   - _Depends: 2.3, 3.2_
   - _Requirements: 2.1, 2.2, 2.4, 8.1, 8.2_
+  - _Verified 2026-09-13:_ independent review APPROVED after an unknown-call boundary repair; fresh suite 1,128 passed, nine skipped; global Ruff/format/mypy passed. Real PHP/JavaScript Git guard-removal snapshots reached unchanged helper-file witnesses with local dirty state preserved. Production negative regression reproduced then fixed unknown(req.body) being pruned as tier zero: modeled arguments no longer summarize unknown consumers. Existing ranking and detector semantics unchanged; unsupported projections remain explicit. Evidence: `benchmarks/measurements/2026-09-13-change-context-smoke.json`. Lexical attribution is not semantic novelty or alert admission.
 - [ ] 3.4 Apply vendor exclusion and explicit language partitions
   - Reuse declared layout exclusions before target enumeration and graph build; retain unshipped first-party tests and separate browser/server or language capability coverage.
   - Build supported frontend partitions explicitly; preserve unknown cross-language and vendor semantics without restoring excluded sources.