claude-token-saver 3.2.0 → 3.2.2

package/README.en.md

@@ -251,6 +251,12 @@ Also update `statusLine.command` in `~/.claude/settings.json` to `claude-token-s
 
 ## Release notes
 
+### v3.2.2 (2026-07-13)
+- **No re-proposing registered rules** — (tier|category|project) combinations that already have a model-fitting rule are excluded from scan candidates (a global rule covers every project). Fixes rules registered outside the promote flow (e.g. migrations) resurfacing as candidates.
+
+### v3.2.1 (2026-07-13)
+- **Instant pattern analysis on first setup** — `install` (including npm postinstall) analyzes your existing session logs inline when no cache exists, so tier-delegation candidates appear from the very first Claude Code session (previously the second).
+
 ### v3.2.0 (2026-07-13)
 - **Tier classification (T0/T1/T2)** — route-scan grows from a binary easy/other split into three tiers. New signals: mutating tool calls and tool errors; output thresholds auto-calibrate to the user's own distribution (clamped); a dedicated category for pasted screen/log Q&A; conversational episodes (<100 output tokens) excluded. Design and research evidence in `docs/TIER_CRITERIA.md`.
 - **Model-fitting ratchet separated** — promoted delegation rules live in their own file (`.claude/ratchet-model.md` / `~/.claude/ratchet-model.md`), file-level-separated from hand-written rules, managed via `route-scan rules [rm <N>]`; the harness CLAUDE.md block references both files.

package/README.md

@@ -208,6 +208,12 @@ npm uninstall -g claude-cache-monitor && npm i -g claude-token-saver
 
 ## 릴리스 노트
 
+### v3.2.2 (2026-07-13)
+- **등록 룰 재제안 차단** — 이미 모델 피팅 룰이 있는 (티어|카테고리|프로젝트) 조합은 스캔 후보에서 자동 제외 (글로벌 룰은 전 프로젝트 커버). promote를 거치지 않고 등록된 룰(마이그레이션 등)이 후보로 되살아나던 문제 수정.
+
+### v3.2.1 (2026-07-13)
+- **최초 설치 시 즉시 패턴 분석** — `install`(npm postinstall 포함)이 캐시가 없으면 기존 세션 로그를 그 자리에서 분석해, 첫 Claude Code 세션부터 티어 위임 후보가 표시됩니다 (기존에는 두 번째 세션부터).
+
 ### v3.2.0 (2026-07-13)
 - **티어 분류 (T0/T1/T2)** — route-scan이 이분법(easy/그외)에서 3티어로 진화. 신호에 변경성 도구 수·도구 에러 수 추가, 출력 임계값은 사용자 분포 기반 자동 보정(클램프 포함), 붙여넣은 화면·로그 질문 전용 카테고리 신설, 대화성 응답(출력 <100토큰) 제외. 기준 설계·리서치 근거는 `docs/TIER_CRITERIA.md`.
 - **모델 피팅 랫쳇 분리** — 승격된 위임 룰은 별도 파일(`.claude/ratchet-model.md` / `~/.claude/ratchet-model.md`)에 저장돼 사용자 룰과 파일 단위로 분리. `route-scan rules [rm <N>]`로 관리하며, 하네스 CLAUDE.md 블록이 두 파일을 함께 참조.

package/bin/cli.js

@@ -412,6 +412,25 @@ async function main() {
       print('legacy /token-monitor', r.legacy);
       console.log('  (consolidated into the skill — same workflow, triggered by intent)');
     }
+    // First-time setup: analyze existing session logs right away so the
+    // very first session already sees delegation candidates — without this,
+    // the initial scan would only start from the first session's hook and
+    // its results would surface one session late. Runs inline (a few
+    // seconds on a typical 14-day history): a detached child can be reaped
+    // by sandboxed installers before it finishes, and postinstall carries
+    // `|| true` so a failure here never breaks the install.
+    {
+      const rs = await import('../src/route-scan.js');
+      if (!rs.readRouteScan()) {
+        try {
+          console.log('');
+          console.log('  route-scan: 기존 세션 로그의 사용 패턴을 분석하는 중...');
+          const cache = await rs.runRouteScan({ days: 14 });
+          console.log(`  route-scan: 에피소드 ${cache.totalEpisodes}건 분석 완료 — 위임 후보 ${cache.candidates.length}건.`);
+          console.log('              (다음 Claude Code 세션에서 티어 위임 후보가 표시됩니다)');
+        } catch { /* hook-triggered scan covers it on first session instead */ }
+      }
+    }
     console.log('');
     console.log('Open Claude Code in any directory and just mention:');
     console.log('  "cache hit rate" / "1M context" / "5H cap" — the skill auto-activates.');

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-token-saver",
-  "version": "3.2.0",
+  "version": "3.2.2",
   "description": "Save tokens on Claude Code — spike diagnosis, 1M-context detection, TTL countdown, statusline. (formerly claude-cache-monitor)",
   "type": "module",
   "bin": {

package/src/route-scan.js

@@ -285,12 +285,25 @@ export async function runRouteScan({ days = 14 } = {}) {
   const prev = readRouteScan();
   const resolved = new Set(prev?.resolved || []);
 
+  // Never re-propose a pattern that already has a registered model-fitting
+  // rule (a global rule covers the category in every project). Without this,
+  // rules that entered the registry outside the promote flow — migrations,
+  // future imports — would resurface as candidates forever.
+  let registered = [];
+  try {
+    const { loadModelRules } = await import('./model-rules.js');
+    registered = loadModelRules().rules;
+  } catch { /* registry unreadable — candidates may repeat until promote */ }
+  const hasRule = (g) => registered.some((r) =>
+    r.tier === g.tier && r.category === g.category &&
+    (r.scope === 'global' || r.project === g.project));
+
   const ruleText = (g) => g.tier === 'T2'
     ? `"${g.label}" 유형의 단순 요청(예: "${g.example}")은 ${g.agent}(haiku) 서브에이전트로 위임한다`
     : `"${g.label}" 유형의 중간 난도 요청(예: "${g.example}")은 model: sonnet 서브에이전트로 위임한다 (설계 판단·반복 에러 발생 시 메인 모델이 이어받음)`;
 
   const candidates = [...groups.values()]
-    .filter((g) => g.count >= MIN_RECURRENCE)
+    .filter((g) => g.count >= MIN_RECURRENCE && !hasRule(g))
     .sort((a, b) => b.count - a.count)
     .slice(0, 8)
     .map((g, i) => ({