diff --git a/packages/codev/src/__tests__/prompt-behavior-metrics.test.ts b/packages/codev/src/__tests__/prompt-behavior-metrics.test.ts
index d422764ee..b066838e5 100644
--- a/packages/codev/src/__tests__/prompt-behavior-metrics.test.ts
+++ b/packages/codev/src/__tests__/prompt-behavior-metrics.test.ts
@@ -176,9 +176,40 @@ describe('behavioural metrics (M12a / T14)', () => {
     }
   });
 
+  // The exact project set the committed baseline (n=160) was computed over.
+  // FROZEN by definition: the baseline is a point-in-time artifact, so its
+  // reproduction must pin its sample. Do NOT add new projects here to make a
+  // failure go away — a change in these numbers over THIS set means the
+  // measurement code's semantics changed, which is what this test guards.
+  const BASELINE_SAMPLE = [
+    '0092-terminal-file-links',
+    '0120-codex-sdk-integration',
+    '0124-test-suite-consolidation',
+    '403-af-send-typing-awareness',
+    '456-dashboard-statistics-tab-in-ri',
+    '462-add-spike-protocol-for-technic',
+    '467-add-open-files-shells-section-',
+    '468-af-rename-command-to-rename-cu',
+    '589-support-non-github-repositorie',
+    '723-improve-arch-md-lessons-learne',
+    '746-spir-architect-s-baked-archite',
+    '755-multi-architect-support-per-ar',
+    '761-surface-multiple-architects-in',
+    '778-gemini-cli-antigravity-cli-jun',
+    '786-multi-architect-feature-is-und',
+    '823-multi-architect-coordination-b',
+    '927-needs-attention-surface-prs-vi',
+    '987-engineering-wisdom-is-write-on',
+  ];
+
   it('reproduces the committed baseline numbers on this repo', () => {
     // Guards against a refactor silently changing what the baseline means.
-    const m = measureBehavior(REPO_ROOT);
+    // Pinned to the baseline's own sample so projects that post-date the
+    // baseline (each adding review verdicts to codev/projects/) cannot
+    // perturb the reproduction — the un-pinned form broke the moment any
+    // later project ran its first consultation (160 → 163 via project 1286).
+    const m = measureBehavior(REPO_ROOT, { includeProjects: BASELINE_SAMPLE });
+    expect(m.sampleProjects).toEqual(BASELINE_SAMPLE);
     expect(m.b1_totalVerdicts).toBe(160);
     expect(m.b1_requestChangesRate).toBeCloseTo(51.88, 1);
     expect(m.b2_roundsPerPhaseMean).toBeCloseTo(1.12, 2);
diff --git a/packages/codev/src/lib/prompt-behavior-metrics.ts b/packages/codev/src/lib/prompt-behavior-metrics.ts
index e325c440f..11507adfc 100644
--- a/packages/codev/src/lib/prompt-behavior-metrics.ts
+++ b/packages/codev/src/lib/prompt-behavior-metrics.ts
@@ -158,11 +158,12 @@ function walk(dir: string, pred: (p: string) => boolean): string[] {
  * review history for protocols that loop that way, so pir/bugfix/air contribute
  * nothing. Sorted for deterministic output.
  */
-function collectReviewMetrics(root: string, excludeProjects: string[]) {
+function collectReviewMetrics(root: string, excludeProjects: string[], includeProjects?: string[]) {
   const projDir = path.join(root, 'codev', 'projects');
   const files = fs.existsSync(projDir)
     ? fs
         .readdirSync(projDir)
+        .filter((d) => (includeProjects ? includeProjects.includes(d) : true))
         .filter((d) => !excludeProjects.includes(d))
         .map((d) => path.join(projDir, d, 'status.yaml'))
         .filter((f) => fs.existsSync(f))
@@ -264,6 +265,18 @@ export interface MeasureOptions {
    * test failed the moment this project's own iter-1 review landed).
    */
   excludeProjects?: string[];
+  /**
+   * Project directory names to measure EXCLUSIVELY (an allowlist), applied
+   * before `excludeProjects`. Unset means "every project on disk".
+   *
+   * A frozen baseline needs a frozen sample: the committed baseline artifact
+   * was computed over the projects that existed at baseline time, so a
+   * reproduction against live history must pin that set — otherwise every
+   * subsequent project that runs consultations perturbs the numbers and the
+   * reproduction fails for reasons that have nothing to do with the metrics
+   * (discovered when project 1286's first review iteration moved B1 160 → 163).
+   */
+  includeProjects?: string[];
   /**
    * Basename prefixes excluded from B3's prose scan, for the same reason: this
    * project's own thread/review discuss scar rules at length and grow with
@@ -278,7 +291,7 @@ export const SELF_FILE_PREFIXES = ['spir-1252_', '1252-'];
 
 /** Collect all behavioural metrics for a repo root. Deterministic over B1–B4. */
 export function measureBehavior(root: string, opts: MeasureOptions = {}): BehaviorMetrics {
-  const rv = collectReviewMetrics(root, opts.excludeProjects ?? [SELF_PROJECT_DIR]);
+  const rv = collectReviewMetrics(root, opts.excludeProjects ?? [SELF_PROJECT_DIR], opts.includeProjects);
   const scar = collectScarHits(root, opts.excludeFilePrefixes ?? SELF_FILE_PREFIXES);
   return {
     b1_requestChangesRate: round2(rv.rcRate * 100),