Apply B1+B2+B3+B4 from 2026-04-30 cross-lineage scrum

Cross-lineage scrum (Opus 4.7 / Kimi K2.6 / Qwen3-coder via chatd's
/v1/chat) on the harness's first 4 commits surfaced 5 real bugs;
this commit lands the 4 inside the LLM/validator stack. B5 (scanner
skip-list semantics) ships separately as it changes scan behavior
on every target repo.

B1 (Kimi BLOCK + Opus WARN convergent) — internal/validators:
evidencePresent had two flaws: (1) cursor advanced on match in the
trim-line fallback, breaking same-line repeated matches AND skipping
not-yet-considered lines so out-of-order evidence spuriously failed;
(2) strings.Contains on a single `}` trim-matched any closing brace
in the file, defeating the "evidence quotes real text" contract.
Fix: trivial-evidence guard FIRST (reject anything <4 non-whitespace
chars) + per-line search no longer advances a cursor. New regression
test TestEvidencePresent_RejectsTrivialMatches covers `}`, `{`, `)`,
empty, and out-of-order multi-line evidence (which now passes —
order isn't part of the contract).

B2 (Kimi WARN + Opus WARN convergent) — internal/pipeline:
WriteJSON error for rejected-findings.json was swallowed with
`if err == nil`, so a write failure left the validation phase
reporting status="ok" while the audit trail vanished. Mirror the
validated-findings branch: surface the error in
validatePhase.Errors + bump status to degraded + ExitCode=66.

B3 (Kimi BLOCK + Opus BLOCK convergent) — internal/llm/ollama.go:
HealthCheck.basic_prompt_ok was set to true on ANY non-empty
response, so a model emitting `<think>...` traces or apologies
passed silently. Now requires the response to contain "OK"
(uppercase, substring). Substring rather than equality lets minor
whitespace/punctuation variations through (some models add a
trailing period). Errors now record what the model actually said
when it fails the check.

B4 (Opus BLOCK only — same class as today's chatd Anthropic-temp
fix) — internal/llm/ollama.go: chatBody had `if opts.Temperature != 0`
which silently dropped Temperature=0 from the request, so HealthCheck
+ Reviewer (both pass Temperature=0 expecting determinism) actually
ran at Ollama's ~0.8 default. Always forward Temperature now. The
two callers always set explicit values, so "0 means 0" is correct;
if a future caller wants Ollama's default they'll switch
CompleteOptions.Temperature to *float64 like chatd did this morning.

Verified end-to-end: insecure-repo + --enable-llm still produces 25
confirmed findings (16 static + 9 LLM), 0 rejected. Validator unit
tests: 11 pass (added TestEvidencePresent_RejectsTrivialMatches).

Same-day-as-shipping scrum, same-day-as-shipping fixes. The
convergent-≥2 gate caught 3 of these; the 4th was Opus-only but
verified by reading the code (same idiom as today's chatd bug).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

internal/validators/validate.go

@@ -178,35 +178,57 @@ func check(repoPath string, f analyzers.Finding, cache map[string]string) Result
 }
 
 // evidencePresent returns true if the evidence appears verbatim in
-// the file. Multi-line evidence is checked as-is; single-line evidence
-// is also compared trimmed (models often add/drop leading whitespace
-// when quoting code).
+// the file, OR every evidence line trim-matches some line in the
+// file (models often re-indent quotes when quoting code).
+//
+// Scrum fix B1 (Kimi BLOCK + Opus WARN, 2026-04-30):
+//   - reject trivially-matchable evidence FIRST (empty, lone braces,
+//     single-char/punct quotes); strings.Contains on tiny strings
+//     hits half the file and lets a non-quoting LLM pass
+//   - per-line trim-match no longer advances a cursor on hit; earlier
+//     version drove cursor forward unconditionally, both preventing
+//     same-line repeated matches and skipping unseen lines, so
+//     out-of-order evidence spuriously failed
+//
+// Order is no longer enforced — every evidence line just needs to
+// appear somewhere in the file. The contract is "evidence quotes
+// real text from the file," not "evidence quotes contiguous text in
+// the same order."
 func evidencePresent(content, evidence string) bool {
+	trimmed := strings.TrimSpace(evidence)
+	if trimmed == "" {
+		return false
+	}
+	// Trivial-match guard: if the *entire* evidence is shorter than 4
+	// non-whitespace chars, reject regardless of how it's being matched.
+	// Lone `}` / `{` / `)` / `(` would substring-hit any matching brace
+	// in the file. Min-length picked at 4 because real verbatim quotes
+	// (variable names, function calls) are essentially always longer.
+	if nonWSLen(trimmed) < 4 {
+		return false
+	}
 	if strings.Contains(content, evidence) {
 		return true
 	}
-	// Trim each line of the evidence + match line-by-line. Conservative:
-	// every evidence line must appear (in order) in the file's trimmed
-	// lines for the evidence to count as found.
-	evLines := strings.Split(strings.TrimSpace(evidence), "\n")
-	if len(evLines) == 0 {
-		return false
-	}
+	evLines := strings.Split(trimmed, "\n")
 	contentLines := strings.Split(content, "\n")
-	cursor := 0
 	for _, ev := range evLines {
 		want := strings.TrimSpace(ev)
 		if want == "" {
 			continue
 		}
+		// Per-line trivial guard: even within a multi-line evidence
+		// block, a line of `}` shouldn't satisfy the "this evidence
+		// line appears in the file" check.
+		if nonWSLen(want) < 4 {
+			return false
+		}
 		found := false
-		for cursor < len(contentLines) {
-			if strings.Contains(strings.TrimSpace(contentLines[cursor]), want) {
+		for _, cl := range contentLines {
+			if strings.Contains(strings.TrimSpace(cl), want) {
 				found = true
-				cursor++
 				break
 			}
-			cursor++
 		}
 		if !found {
 			return false
@@ -215,6 +237,16 @@ func evidencePresent(content, evidence string) bool {
 	return true
 }
 
+func nonWSLen(s string) int {
+	n := 0
+	for _, r := range s {
+		if r != ' ' && r != '\t' && r != '\n' && r != '\r' {
+			n++
+		}
+	}
+	return n
+}
+
 // highestLine extracts the largest line number cited in the hint.
 // Accepts "42", "10-20" (returns 20), "line 42", "L42", "42:5".
 // Returns (n, true) on parse; (0, false) if no number found.