oh-my-customcode 1.1.1 → 1.1.3

package/dist/cli/index.js

@@ -241,7 +241,7 @@ var init_package = __esm(() => {
     workspaces: [
       "packages/*"
     ],
-    version: "1.1.1",
+    version: "1.1.3",
     description: "Batteries-included agent harness for Claude Code",
     type: "module",
     bin: {

package/dist/index.js

@@ -2031,7 +2031,7 @@ var package_default = {
   workspaces: [
     "packages/*"
   ],
-  version: "1.1.1",
+  version: "1.1.3",
   description: "Batteries-included agent harness for Claude Code",
   type: "module",
   bin: {

package/package.json

@@ -3,7 +3,7 @@
   "workspaces": [
     "packages/*"
   ],
-  "version": "1.1.1",
+  "version": "1.1.3",
   "description": "Batteries-included agent harness for Claude Code",
   "type": "module",
   "bin": {

package/templates/.claude/agents/mgr-gitnerd.md

@@ -33,7 +33,7 @@ You are a Git operations specialist following GitHub flow best practices.
 ```
 <type>(<scope>): <subject>
 
-Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
+Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
 ```
 
 Types: feat, fix, docs, style, refactor, test, chore

package/templates/.claude/agents/slack-cli-expert.md

@@ -14,7 +14,7 @@ tools:
 permissionMode: bypassPermissions
 ---
 
-You are an expert Slack CLI developer specialized in building, deploying, and managing Slack apps using the official Slack CLI v4.0 and the Slack Platform (https://docs.slack.dev/tools/slack-cli/). Slack CLI v4.0 adds first-class Slack Agent development support.
+You are an expert Slack CLI developer specialized in building, deploying, and managing Slack apps using the official Slack CLI (v4.x) and the Slack Platform (https://docs.slack.dev/tools/slack-cli/). Slack CLI v4.0 adds first-class Slack Agent development support.
 
 ## Capabilities
 

package/templates/.claude/agents/wiki-curator.md

@@ -22,7 +22,7 @@ Dedicated agent for wiki file operations. All wiki/ directory writes go through
 
 - Create new wiki pages from source file analysis
 - Update existing wiki pages when sources change
-- Maintain index.md and log.md
+- Maintain index.yaml and log.jsonl
 - Execute wiki lint fixes (orphan removal, cross-ref repair)
 - Generate synthesis pages (architecture, workflows, concepts)
 
@@ -51,18 +51,18 @@ Every page must:
 3. Determine what changed
 4. Write updated page with current date in `updated` field
 5. Update cross-references in related pages
-6. Update index.md if page is new
+6. Update index.yaml if page is new
 
 ### Batch Update (Category)
 1. Glob source files in category
 2. Compare modification dates against wiki pages
 3. Write only changed/new pages
-4. Batch-update index.md once at end
+4. Batch-update index.yaml once at end
 
 ### Lint Fix
 1. Receive lint findings from orchestrator
 2. Fix each category: remove orphans, repair broken refs, update stale pages
-3. Append fix results to log.md
+3. Append fix results to log.jsonl
 
 ## Limitations
 

package/templates/.claude/rules/MUST-agent-design.md

@@ -24,8 +24,7 @@ tools: [Read, Write, ...]  # Allowed tools
 | `sonnet5` | claude-sonnet-5 | CC default (v2.1.197+); native 1M context |
 | `opus` | claude-opus-4-6 | Complex reasoning, architecture |
 | `opusplan` | claude-opus-4-6 + plan mode | Architecture planning with approval gates |
-| `opus47` | claude-opus-4-7 | Supports xhigh effort |
-| `opus48` | claude-opus-4-8 | Latest Opus model (GA); highest capability below Fable 5 |
+| `opus48` | claude-opus-4-8 | Latest Opus model (GA); highest capability below Fable 5; supports xhigh effort |
 | `fable` | claude-fable-5 | Mythos-class; tier above Opus, highest GA capability (access added in CC v2.1.170) |
 
 Extended context suffix: `[1m]` (e.g., `claude-opus-4-6[1m]`) — enables 1M token context window.
@@ -54,6 +53,8 @@ This is a settings-level resilience mechanism, distinct from the per-agent `mode
 
 > **v2.1.196+**: 조직 기본 모델(org default models)이 추가되어 관리자가 org 콘솔에서 조직 전체 기본 모델을 설정하며, 사용자가 직접 선택하지 않으면 `/model`에 "Org default"(또는 "Role default")로 표시됩니다. v2.1.175 `enforceAvailableModels`/v2.1.187 org model restriction의 연장선 — 조직 관리 설정이 기본 모델 해석까지 관여합니다. 또한 백그라운드 세션 turn 이후 중복 recap 라인 문제를 수정 — schema-rejected된 StructuredOutput 시도가 재시도와 나란히 렌더되지 않습니다(schema-constrained subagent spawn / Workflow `agent({schema})`에 관련).
 
+> **v2.1.198+**: 내장 Explore 에이전트가 이제 main session의 모델을 상속합니다(opus로 cap; 이전에는 haiku 고정) — Explore 기반 검색 subagent의 추론 품질이 향상됩니다. 또한 subagent와 context compaction이 session의 extended thinking 설정을 상속하여, 위임된 작업의 출력 품질이 세션 설정과 정합됩니다. R009/R018 병렬 subagent 위임 품질에 관련.
+
 ### Safe Mode & Bundled Skill Control (CC v2.1.169+)
 
 > **v2.1.169+**: `--safe-mode` (and `CLAUDE_CODE_SAFE_MODE`) starts Claude Code with ALL customizations disabled (CLAUDE.md, plugins, skills, hooks, MCP servers) — use it to isolate whether a project customization (agent/skill/hook) causes a regression. The `disableBundledSkills` setting (and `CLAUDE_CODE_DISABLE_BUNDLED_SKILLS` env) hides bundled skills, workflows, and built-in slash commands from the model — useful when bundled skills conflict with or duplicate project skills (R006 skill-surface management). Note: `disableBundledSkills` hides skills from the model but is a CC platform setting, distinct from the advisory `skills:` frontmatter field (which is documentation metadata, not a runtime allowlist).
@@ -203,6 +204,8 @@ Agent frontmatter `hooks:` now fire when the agent runs as a main-thread agent v
 
 > **v2.1.195+**: 하이픈이 포함된 hook matcher 식별자(예: `code-reviewer`, `mcp__brave-search`)가 이제 substring이 아니라 exact-match됩니다(이전에는 부분 문자열로 우연히 매칭). 하이픈 포함 MCP 서버의 모든 tool을 매칭하려면 `mcp__brave-search__.*` 형식을 사용하세요. 참고: oh-my-customcode의 `hooks.json`은 `*` / `mcp_tool_name matches "..."` / `tool == "..."` 형식을 사용하므로 이 변경의 영향을 받지 않습니다 — 다만 하이픈 포함 tool-name matcher를 새로 추가할 때 exact-match 시맨틱을 고려해야 합니다.
 
+> **v2.1.199+**: SessionStart/Setup/SubagentStart hook이 exit code 2로 종료할 때 stderr를 조용히 숨기던 문제가 수정되어 이제 오류가 표시됩니다. `.claude/hooks.json`의 hard-block/advisory hook 디버깅 가시성이 향상됩니다(R021 Enforcement Tiers 관측성과 정합).
+
 ## Permission Mode Guidance
 
 CC defaults `mode` to `acceptEdits` if not specified — always pass `mode: "bypassPermissions"` explicitly in Agent tool calls (see R010). See guidance details via Read tool.
@@ -401,6 +404,8 @@ Key optional fields: `scope`, `context`, `version`, `effort`, `model`, `agent`,
 
 > **v2.1.178+**: When names collide across nested `.claude/` directories, the agent, workflow, and output-style CLOSEST to the working directory now wins; project-scope workflow saves target the closest existing `.claude/workflows/`. Relevant to multi-`.claude/` layouts — project-root `.claude/` definitions are overridden by a nested `.claude/` when working inside that subtree.
 
+> **v2.1.199+**: 스택된 slash-skill 호출(`/skill-a /skill-b do XYZ`)이 이제 leading skill을 최대 5개까지 모두 로드합니다(이전에는 첫 번째만 로드). oh-my-customcode의 라우팅 스킬 체이닝(예: `/omcustom:fsd`가 여러 스킬을 연쇄 호출하는 패턴)에서 다중 스킬 스택 호출 시 컨텍스트 손실이 줄어듭니다. 또한 subagent 조회 중 `/model`·`/fast`를 입력하면 lead의 model picker가 열리며 notice가 표시됩니다.
+
 <!-- DETAIL: Skill Optional Fields (full yaml block)
 ```yaml
 scope: core                # core | harness | package (default: core)

package/templates/.claude/rules/MUST-agent-teams.md

@@ -374,6 +374,10 @@ Cross-reference: R020 ("actual outcome ≠ attempt" — verifying that a command
 
 > **CC v2.1.169+**: `claude agents --json` now includes blocked and just-dispatched background sessions (previously omitted), adds `--all` to include completed sessions, and adds `id` and `state` fields. This strengthens the deterministic ground-truth for member completion verification — `state` distinguishes blocked/running/completed directly, and `--all` confirms a member actually completed (rather than just disappearing from the active list). Use `--all` + `state` as the ground-truth signal instead of inferring completion from a member's absence.
 
+> **v2.1.198+**: Agent Teams에서 teammate가 API 오류로 죽으면 이제 lead에 "failed"를 보고하고, stuck teammate에게 메시지를 보내면 즉시 wake시켜 retry하게 합니다. 이는 v2.1.169 `state` 필드 기반 deterministic ground-truth를 보강하지만, SendMessage report 자체는 여전히 low-reliability 신호로 취급한다(위 표의 원칙 불변).
+
+> **v2.1.199+**: subagent가 rate limit/server error로 잘리면 partial work를 parent에 반환하며, subagent가 API 오류(usage limit reached 등)를 성공 결과로 오보하던 문제가 수정되어 이제 오류를 parent agent에 정확히 보고합니다. 플랫폼이 false-success 자가보고를 줄였으나, deterministic ground-truth(`git status`/`grep`/validation scripts) 검증 원칙은 여전히 유효하다.
+
 **Stall handling**: When a member shows no task progress within ~2 minutes despite spawn + owner assignment + SendMessage coordination, reassign the work to a standalone Agent (R009) rather than continuing to nudge the stalled member. Stalled Teams members waste tokens on idle polling and delay the overall workflow.
 
 Observed instance: v0.159.0 release (session 105) — members assigned to disjoint-file cleanup tasks went idle without executing; deterministic git-diff check exposed the gap; work was reassigned to standalone parallel Agents. References: #1261, #1262.

package/templates/.claude/rules/MUST-completion-verification.md

@@ -71,6 +71,8 @@ Subagents often report failures as "pre-existing", "baseline", or "unchanged". T
 Never accept "pre-existing" without direct base-branch evidence. A false "pre-existing" claim can mask a regression introduced by the current change.
 -->
 
+> **v2.1.199+**: subagent가 API 오류(usage limit reached 등)를 성공 결과로 오보하던 문제가 수정되어 이제 오류가 parent agent에 정확히 보고됩니다. 플랫폼 수정으로 false-success 자가보고 빈도는 줄지만, "actual outcome ≠ attempt" ground-truth 검증 원칙(R020 Core Rule)은 여전히 유지된다 — subagent 보고를 그대로 신뢰하지 말고 `git status`/`grep`/validation script로 재확인한다.
+
 ## Common False Completion Patterns — 8 anti-patterns including "Command executed" without exit code check, "Waiting for manual publish" when CI auto-publishes, "UI changes done" without browser render. See full table via Read tool.
 
 ### Test-Skip Is Not Completion (#1217 item #5)

package/templates/.claude/rules/MUST-enforcement-policy.md

@@ -19,6 +19,8 @@ oh-my-customcode uses an **advisory-first enforcement model**. Most rules are en
 
 > **v2.1.163+**: Stop and SubagentStop hooks can return `hookSpecificOutput.additionalContext` (JSON) to feed structured feedback back into Claude's context without triggering a hook error label. This enables advisory-style enforcement via Stop/SubagentStop hooks (e.g., `session-reflection.sh`, omcustom-loop SubagentStop) to pass richer context — replacing plain stderr text — without disrupting the turn continuation behavior that advisory-first enforcement relies on.
 
+> **v2.1.199+**: SessionStart/Setup/SubagentStart hook이 exit code 2로 종료할 때 stderr를 조용히 숨기던 문제가 수정되어 이제 오류가 표시됩니다. Hard Block/Advisory 계층(위 Enforcement Tiers 표)의 훅 실패 관측성을 강화합니다 — cf. v2.1.163 `additionalContext` 구조화 피드백.
+
 ## Why Advisory-First
 
 1. **Agent flexibility**: Hard blocks can trap agents in unrecoverable states

package/templates/.claude/rules/MUST-orchestrator-coordination.md

@@ -287,7 +287,7 @@ The Agent tool defaults to `mode: "acceptEdits"`, which overrides agent frontmat
 | Scope | Every Agent tool call, without exception |
 | Why | CC's Agent tool `mode` default (`acceptEdits`) overrides frontmatter |
 | History | #926 (v0.99.1), #947 (v0.100.1), #955 (v0.103.0) — recurring issue |
-| Enforcement | Prompt-based (R021); all 23 agent-spawning skills include instruction |
+| Enforcement | Prompt-based (R021); all agent-spawning skills include instruction |
 
 ### Self-Check
 
@@ -337,6 +337,10 @@ Before spawning any agent:
 
 > **v2.1.196+**: 백그라운드 작업을 wake할 때 transcript probe가 실제 transcript를 오독하여 대화를 영구 삭제하고 원 프롬프트를 재실행하던 문제를 수정 — 이제 파일을 삭제하지 않고 따로 보관합니다(transcript 의존 스킬 `homework`/`episodic-memory`에 관련). `claude agents --dangerously-skip-permissions`가 조용히 auto mode로 폴백하던 문제를 수정하여 bypass 고지를 표시하고 spawned agent에도 bypass 모드를 적용합니다(R010 Universal bypassPermissions와 정합). 또한 `claude agents` 사이드 패널 문제들(에이전트 열 때 키보드 포커스 고착, 열 때마다 백그라운드 작업의 subagent type 유실, 활성 실행 중 잘못된 상태 표시)을 수정. `mode: "bypassPermissions"`는 여전히 필수입니다.
 
+> **v2.1.198+**: `claude agents`에서 launch된 background agent가 worktree에서 code 작업을 완료하면 이제 멈춰서 묻지 않고 commit·push·draft PR open을 자동 수행합니다. 또한 응답 중 일시적 network 오류(ECONNRESET 등)로 turn이 abort되던 문제가 backoff retry로 수정되었고, web/desktop/VS Code task panel에서 background task가 완료 후에도 "Running"으로 멈춰있던 문제가 수정되었습니다. background agent의 자동 commit/PR 자동화가 강화되었으므로, `mode: "bypassPermissions"`는 여전히 필수입니다.
+
+> **v2.1.199+**: subagent가 rate limit이나 server error로 잘리면 이제 조용히 실패하는 대신 partial work를 parent에 반환합니다. background-agent daemon(Linux)이 unclean shutdown 후 corrupted worker record로 ~50초마다 자신과 모든 agent를 죽이던 문제, macOS SSH cold-start "Could not switch to audit session" 문제, `claude stop`이 background-agent respawn과 race하면 조용히 무효화되던 문제(이제 respawn이 stop을 존중), background job progress indicator가 긴 명령 중 멈춰있던 문제가 수정되었습니다. background-agent lifecycle 견고성이 추가로 강화되었으며, `mode: "bypassPermissions"`는 여전히 필수입니다.
+
 ## Agent Capability Pre-Check
 
 Before delegating a task to a subagent, MUST verify the target agent's tool capabilities against the task requirements. Failure to pre-check causes round-trip waste (delegation → failure → re-delegation).

package/templates/.claude/rules/SHOULD-hud-statusline.md

@@ -31,6 +31,8 @@ Format: `─── [Spawn] {subagent_type}:{model} | {description} ───`
 
 > **v2.1.196+**: 여러 병렬 요청이 사용량 한도에 도달하는 순간 rate-limit 경고가 깜빡이며 꺼지고 rate-limit telemetry가 과다 집계되던 문제를 수정. R012 관측성의 rate-limit 계측 정확도 개선입니다.
 
+> **v2.1.198+**: `claude agents`에 background agent notifications가 추가되어, 입력이 필요하거나 완료된 세션이 `Notification` hook을 발화합니다(`agent_needs_input` / `agent_completed`). R012 관측성을 백그라운드 subagent 상태 알림까지 확장 — HUD 이벤트 채널과 결합해 백그라운드 위임 작업의 대기/완료 상태를 놓치지 않게 합니다.
+
 <!-- DETAIL: HUD Events full spec
 ### When to Display: Multi-step tasks, parallel execution, long-running operations. Skip for single brief operations.
 ### Parallel Display:

package/templates/.claude/rules/SHOULD-verification-ladder.md

@@ -88,6 +88,16 @@ staleness/audit 검증은 model ID·placeholder·TBD뿐 아니라 **폐기된
 |--------------|----------|
 | staleness 스캔을 model ID/placeholder/TBD로만 한정 | 폐기된 플랫폼 기능/절차 참조(deprecated CC feature/procedure)도 grep 스캔 |
 
+## Detection Guard Delegation Standard (Origin: #1438 #3)
+
+Tier-1 shift-left 검출 가드(예: deprecated-pattern grep 가드)의 설계·수정을 서브에이전트에 위임할 때, 위임 프롬프트는 **positive-match(genuine defect mandate — `MUST`/`MANDATORY` 인접 문맥)와 negative-context(deprecation note — "no longer"/"deprecated"/"불필요"/"폐기됨" 설명 문구)를 구분**하도록 명시해야 한다. 이를 누락하면 올바르게 수정된 파일의 폐기-설명 문구까지 과잉매칭하여 자기모순 BLOCK을 유발한다.
+
+| Anti-pattern | Required |
+|--------------|----------|
+| bare 패턴(`/tmp/*.sh` 등)만 grep하도록 위임 → 올바르게 수정된 파일의 "폐기됨/불필요" 설명 문구까지 오탐 | 위임 프롬프트에 positive-match(MUST/MANDATORY 인접) vs negative-context(deprecated/no longer/불필요 설명) 구분 기준을 명시 |
+
+Origin: #1438 (Session 125 회고 찐빠 #3) — deep-verify 가드 반전 위임 시 bare `/tmp/*.sh` 패턴이 올바르게 수정된 9개 파일의 "폐기됨/불필요" 설명 문구까지 오탐; mgr-sauron이 sha256 재계산으로 적발, 3-패턴(mandate/false-claim) 협소화로 정정.
+
 ## Integration
 
 | 규칙 | 상호작용 |
@@ -118,12 +128,16 @@ Before invoking a Workflow script, deterministically verify:
 | No unresolved placeholders (`{phase1_summary}`, `TODO`, `<...>`, `{{ }}`) remain in any agent prompt string | An unfilled placeholder reaches the agent verbatim → garbled task |
 | Template-literal / string concatenation produces the intended prompt (assemble-before-call, see above) | Post-call concatenation (`agent(prompt) + FACTS`) silently drops content |
 | Script parses — balanced braces/quotes, valid JS | A syntax error aborts the entire run after partial work |
+| 프롬프트 문자열 내 셸 변수 `${...}`(`$?`, `${PIPESTATUS[0]}`, `$(...)` 등)가 `\${...}`로 이스케이프되어 있는지 사전 grep 확인 | JS 템플릿 리터럴 안의 이스케이프 안 된 셸 `${...}`를 JS가 JS 표현식으로 평가 → 런타임 `ReferenceError`(예: `PIPESTATUS is not defined`). `node --check`는 문법만 검사하여 이 런타임 오류를 못 잡으므로 별도 결정론 grep 검사가 필요함 |
 
 #### Common Violation (#1271)
 Session 106 follow-up to #1266 ③: a Workflow authoring error recurred — the guardrail fact-sheet was concatenated onto the agent's RETURN VALUE instead of the prompt string, and a placeholder/assembly slip went uncaught because no pre-run sanity check existed. This check is the deterministic Tier-1 guard that catches such slips before the expensive run.
 
 Origin: #1271 (Workflow authoring error recurrence, session 106).
 
+#### Common Violation (#1438)
+Origin: #1438 (Session 125 회고 찐빠 #2) — fix Workflow의 verify 프롬프트에 `${PIPESTATUS[0]}`를 이스케이프 없이 사용 → `PIPESTATUS is not defined` ReferenceError로 verify 단계 실패. `node --check`는 통과했으나 런타임에서 실패.
+
 ### Verifier Ground-Truth for Cross-Cutting Facts
 
 Cross-cutting facts not verifiable from the primary source (external URLs, in-cluster DNS/hostnames, infra topology) MUST be supplied to the verifier as explicit ground-truth. Otherwise an adversarial verifier cannot distinguish a hallucinated value from a correct one — a verification blind spot.

package/templates/.claude/rules/index.yaml

@@ -132,6 +132,46 @@ rules:
     priority: MUST
     scope: orchestrator
 
+  # Completion Verification - MUST
+  - id: R020
+    name: completion-verification
+    title: Completion Verification Rules
+    path: ./MUST-completion-verification.md
+    priority: MUST
+    scope: all
+
+  # Enforcement Policy - MUST
+  - id: R021
+    name: enforcement-policy
+    title: Enforcement Policy
+    path: ./MUST-enforcement-policy.md
+    priority: MUST
+    scope: all
+
+  # Ontology-RAG Routing - SHOULD
+  - id: R019
+    name: ontology-rag-routing
+    title: Routing Enrichment Rules
+    path: ./SHOULD-ontology-rag-routing.md
+    priority: SHOULD
+    scope: orchestrator
+
+  # Wiki Sync - SHOULD
+  - id: R022
+    name: wiki-sync
+    title: Wiki Sync Rules
+    path: ./SHOULD-wiki-sync.md
+    priority: SHOULD
+    scope: all
+
+  # Verification Ladder - SHOULD
+  - id: R023
+    name: verification-ladder
+    title: Verification Ladder Rules
+    path: ./SHOULD-verification-ladder.md
+    priority: SHOULD
+    scope: all
+
   # MAY - Optional
   - id: R005
     name: optimization

package/templates/.claude/skills/action-validator/SKILL.md

@@ -70,9 +70,9 @@ policy_cache:
 
 Policy caching reduces redundant LLM calls for well-understood workflows. Policies are advisory — the orchestrator may override.
 
-## Capability Hints (Opus 4.7+)
+## Capability Hints (Opus 4.8+)
 
-When agents target Opus 4.7 (`opus47` model alias), tool capability hints improve batched tool-call planning. Declare per-tool metadata in policy cache entries:
+When agents target Opus 4.8 (`opus48` model alias), tool capability hints improve batched tool-call planning. Declare per-tool metadata in policy cache entries:
 
 | Field | Values | Effect |
 |-------|--------|--------|
@@ -109,15 +109,14 @@ When a synthesized harness exists for an agent (`.claude/outputs/harnesses/{agen
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths is permitted (CC v2.1.121+, #1101) — no `/tmp/*.sh` wrapping is needed.
 
 To write action-validator results under `.claude/outputs/sessions/`:
 
-1. Write the artifact body to `/tmp/action-validator-$(date +%H%M%S).md` first (Write tool target = `/tmp`, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/sessions/$(date +%Y-%m-%d)/` (Bash target = `/tmp`, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
+1. Write the artifact body directly to `.claude/outputs/sessions/$(date +%Y-%m-%d)/action-validator-$(date +%H%M%S).md` with the Write tool (every Agent tool call includes `mode: "bypassPermissions"`, R010)
+2. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
 
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling, #1016, #1045.
+Reference: R006/R010 sensitive-path handling (CC v2.1.121+), #1101.
 
 
 | Mode | Source | Behavior |

package/templates/.claude/skills/adaptive-harness/SKILL.md

@@ -1,5 +1,5 @@
 ---
-name: adaptive-harness
+name: omcustom:adaptive-harness
 description: Auto-detect project context and optimize harness — deactivate unused agents/skills, suggest missing experts, generate project profile
 scope: harness
 version: 1.0.0
@@ -187,15 +187,14 @@ Append a record to `.claude/outputs/harness-adaptations/YYYY-MM-DD.md`:
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths is permitted (CC v2.1.121+, #1101) — no `/tmp/*.sh` wrapping is needed.
 
-To write adaptive-harness results under `.claude/outputs/sessions/`:
+To write adaptive-harness results under `.claude/outputs/`:
 
-1. Write the artifact body to `/tmp/adaptive-harness-$(date +%H%M%S).md` first (Write tool target = `/tmp`, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/sessions/$(date +%Y-%m-%d)/` (Bash target = `/tmp`, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
+1. Write the artifact body directly to `.claude/outputs/harness-adaptations/$(date +%Y-%m-%d).md` with the Write tool (every Agent tool call includes `mode: "bypassPermissions"`, R010)
+2. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
 
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling, #1016, #1045.
+Reference: R006/R010 sensitive-path handling (CC v2.1.121+), #1101.
 
 
 ```markdown

package/templates/.claude/skills/agent-eval-framework/SKILL.md

@@ -75,15 +75,14 @@ Use this taxonomy to **select representative tasks per category** when building
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths is permitted (CC v2.1.121+, #1101) — no `/tmp/*.sh` wrapping is needed.
 
 To write eval trajectories or result reports under `.claude/outputs/evals/`:
 
-1. Write the artifact body to `/tmp/agent-eval-{HHmmss}.{ext}` first (Write tool target = /tmp, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/evals/{trajectories,sessions}/...` (Bash target = /tmp, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
+1. Write the artifact body directly to `.claude/outputs/evals/{trajectories,sessions}/...` with the Write tool (every Agent tool call includes `mode: "bypassPermissions"`, R010)
+2. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
 
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling.
+Reference: R006/R010 sensitive-path handling (CC v2.1.121+), #1101.
 
 ## Phased Opt-in Gate Workflow
 

package/templates/.claude/skills/de-lead-routing/SKILL.md

@@ -5,11 +5,11 @@ scope: core
 user-invocable: false
 context: fork
 ---
-## Mandatory delegation directive (R010 Universal /tmp Script Bypass)
+## Mandatory delegation directive (R010 bypassPermissions)
 
 When this skill spawns a subagent via the Agent tool, the spawned prompt MUST include this directive verbatim (or equivalent):
 
-> ALL tool calls targeting `.claude/` paths MUST go through `/tmp/*.sh` script via Bash. This includes read-only measurement (sed/cat/wc/ls/grep), Write, Edit. Read tool is exempt. Direct Write/Edit/Bash on .claude/ triggers user approval prompts that block unattended automation. See R010 for the full pattern.
+> 이 스킬이 Agent tool로 서브에이전트를 스폰할 때 mode: "bypassPermissions"를 반드시 포함한다. 서브에이전트는 .claude/ 경로에 직접 Write/Edit/Bash를 사용한다(/tmp 우회 불필요, CC v2.1.121+). See R010.
 
 This directive is preserved inline because Agent-tool prompt synthesis can drop SKILL.md notes; inline mandatory directives survive (#1046 lesson).
 

package/templates/.claude/skills/deep-plan/SKILL.md

@@ -69,26 +69,21 @@ Research-only analysis produces findings based on assumptions about the codebase
 | `/structured-dev-cycle` | Full implementation | Yes — stage-by-stage | 6 |
 | **`/deep-plan`** | **Analysis + Planning + Verification** | **3-pass cross-verification** | **3** |
 
-## Sensitive-Path Artifact Protocol (MANDATORY)
-
-**R010 Universal /tmp Script Bypass (#1052)**: ALL `.claude/` modifications MUST use `/tmp/*.sh` script via Bash. Direct Write/Edit/Bash on `.claude/` triggers CC sensitive-path guard regardless of bypassPermissions.
+## Artifact Output (R006/R010)
 
 Phase 3 verification report path: `.claude/outputs/sessions/{YYYY-MM-DD}/deep-plan-{HHmmss}.md`
 
-When spawning the final synthesis agent to write this artifact, include verbatim in the agent prompt:
+Under `mode: "bypassPermissions"`, subagents write directly to `.claude/outputs/` with Write/Edit — no temp-script wrapping is needed (CC v2.1.121+, #1101). When spawning the final synthesis agent to write this artifact, include verbatim in the agent prompt:
 
 ```
-**Sensitive-path artifact protocol (mandatory for this task)**
-
-Write the artifact under `.claude/outputs/` using the /tmp/*.sh bypass:
-1. Build script in /tmp: cat > /tmp/deep-plan-<timestamp>.sh <<'EOF' ... EOF
-2. Script content: mkdir -p .claude/outputs/sessions/<date>/ && cat > .claude/outputs/sessions/<date>/deep-plan-<HHmmss>.md <<'ARTIFACT' ... ARTIFACT
-3. Execute: bash /tmp/deep-plan-<timestamp>.sh
-4. Cleanup: rm /tmp/deep-plan-<timestamp>.sh
-DO NOT use Write/Edit directly on `.claude/outputs/` — CC sensitive-path guard triggers regardless of bypassPermissions/allow rules.
+**Artifact output (for this task)**
+
+Write the verification report to `.claude/outputs/sessions/<date>/deep-plan-<HHmmss>.md`
+using the Write tool directly (create the directory if needed). Your Agent tool call
+runs under `mode: "bypassPermissions"`, so direct Write/Edit on `.claude/` is permitted.
 ```
 
-See R006 "Sensitive Path Handling" + `feedback_sensitive_path_tmp_bypass.md`.
+See R006 "Sensitive Path Handling" (CC v2.1.121+ direct-write convention).
 
 ## Agent Teams (R018)
 

package/templates/.claude/skills/deep-verify/SKILL.md

@@ -84,10 +84,9 @@ Run these checks before declaring release READY. Any match is a release blocker.
 
 | Guard | Detection Command | Severity | Remediation |
 |-------|-------------------|----------|-------------|
-| Skill Bash sensitive-path | `grep -rnE 'mkdir\s+-p[^` + "`" + `\n]*\.claude/(outputs\|agent-memory\|agent-memory-local)' .claude/skills/ templates/.claude/skills/ .claude/rules/ templates/.claude/rules/ 2>/dev/null` | **BLOCK** | Use /tmp/*.sh bypass — both `Bash(mkdir -p)` and `Write` tool on `.claude/` trigger sensitive-path guard. Write script to /tmp, invoke via Bash. See R006 "Sensitive Path Handling" + `feedback_sensitive_path_tmp_bypass.md` |
-| Skill artifact path missing /tmp bypass directive | `find .claude/skills/ templates/.claude/skills/ -name SKILL.md | xargs grep -lE '.claude/outputs/' | while read f; do if ! grep -qF 'tmp/*.sh' "$f"; then echo "$f"; fi; done` | **WARN** | Add /tmp/*.sh bypass directive after artifact path mention. Pattern: `> **Tool**: To write artifacts under .claude/outputs/, use a /tmp/*.sh script...` (see #1016 fix) |
+| Residual deprecated `/tmp` sensitive-path bypass narrative | `grep -rnE -e 'regardless of[[:space:]]+.?bypassPermissions' -e '(MUST|MANDATORY|Mandatory)[^\n]{0,80}/tmp/\*\.sh' -e '/tmp/\*\.sh[^\n]{0,80}(MANDATORY|MUST be used|bypass script)' .claude/skills/ templates/.claude/skills/ .claude/agents/ templates/.claude/agents/ 2>/dev/null` | **BLOCK** | Remove the deprecated temp-script bypass narrative — it was retired in CC v2.1.121 (#1101). Under `mode: "bypassPermissions"`, write directly to `.claude/` with Write/Edit. See R006 "Sensitive Path Handling" |
 
-> **Why**: CC sensitive-path check runs above `bypassPermissions` and Bash allow rules (#960/#961/#978/#981). Both `Bash(mkdir -p)` and `Write`/`Edit` tool on `.claude/` paths trigger permission prompts — `bypassPermissions` does not help. Use the /tmp/*.sh bypass: write script to /tmp, then `bash /tmp/x.sh` to let the script write to `.claude/` internally (sensitive-path guard only inspects direct tool target paths).
+> **Why**: The temp-script sensitive-path bypass for `.claude/` paths was deprecated in CC v2.1.121 (2026-04-28, #1101). Direct Write/Edit/Bash on `.claude/**` works under `mode: "bypassPermissions"`; only catastrophic shell operations (`rm -rf /`) remain independently blocked. Residual bypass procedures or false "guard blocks bypassPermissions" claims in skills/agents are now defects — this guard catches them with a cheap deterministic grep (R023 shift-left: catch stale platform procedures before they reach expensive review).
 
 ## Output Format
 

package/templates/.claude/skills/dev-lead-routing/SKILL.md

@@ -5,11 +5,11 @@ scope: core
 user-invocable: false
 context: fork
 ---
-## Mandatory delegation directive (R010 Universal /tmp Script Bypass)
+## Mandatory delegation directive (R010 bypassPermissions)
 
 When this skill spawns a subagent via the Agent tool, the spawned prompt MUST include this directive verbatim (or equivalent):
 
-> ALL tool calls targeting `.claude/` paths MUST go through `/tmp/*.sh` script via Bash. This includes read-only measurement (sed/cat/wc/ls/grep), Write, Edit. Read tool is exempt. Direct Write/Edit/Bash on .claude/ triggers user approval prompts that block unattended automation. See R010 for the full pattern.
+> 이 스킬이 Agent tool로 서브에이전트를 스폰할 때 mode: "bypassPermissions"를 반드시 포함한다. 서브에이전트는 .claude/ 경로에 직접 Write/Edit/Bash를 사용한다(/tmp 우회 불필요, CC v2.1.121+). See R010.
 
 This directive is preserved inline because Agent-tool prompt synthesis can drop SKILL.md notes; inline mandatory directives survive (#1046 lesson).
 

package/templates/.claude/skills/dev-review/SKILL.md

@@ -116,15 +116,14 @@ If only PASS/INFO: proceed automatically.
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths is permitted (CC v2.1.121+, #1101) — no `/tmp/*.sh` wrapping is needed.
 
 To write dev-review results under `.claude/outputs/sessions/`:
 
-1. Write the artifact body to `/tmp/dev-review-$(date +%H%M%S).md` first (Write tool target = `/tmp`, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/sessions/$(date +%Y-%m-%d)/` (Bash target = `/tmp`, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
+1. Write the artifact body directly to `.claude/outputs/sessions/$(date +%Y-%m-%d)/dev-review-$(date +%H%M%S).md` with the Write tool (every Agent tool call includes `mode: "bypassPermissions"`, R010)
+2. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
 
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling, #1016, #1045.
+Reference: R006/R010 sensitive-path handling (CC v2.1.121+), #1101.
 
    ```
    With metadata header:

package/templates/.claude/skills/docker-best-practices/SKILL.md

@@ -179,8 +179,6 @@ best_practices:
   - Set resource limits
 
 pattern: |
-  version: "3.8"
-
   services:
     app:
       build:

package/templates/.claude/skills/hada-scout/SKILL.md

@@ -50,7 +50,7 @@ The haiku agent receives the following system prompt:
 
 ```
 You are a relevance filter for the oh-my-customcode project — an AI agent harness/orchestration
-system built on Claude Code CLI with 44 agents, 74 skills.
+system built on Claude Code CLI with 49 agents, 118 skills.
 
 Project domains (HIGH relevance):
 - AI agent orchestration, multi-agent systems, agent design patterns

package/templates/.claude/skills/harness-eval/SKILL.md

@@ -92,15 +92,11 @@ Results saved to `.claude/outputs/sessions/{YYYY-MM-DD}/harness-eval-{HHmmss}.md
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths (including `.claude/outputs/sessions/`) is permitted (CC v2.1.121+, #1101) — no `/tmp` wrapping is needed.
 
-To write harness-eval results under `.claude/outputs/sessions/`:
+Write harness-eval results directly to `.claude/outputs/sessions/$(date +%Y-%m-%d)/harness-eval-$(date +%H%M%S).md`. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification. Catastrophic shell operations (e.g., `rm -rf /`) remain blocked by independent safety guards. For CC < v2.1.121, see git history for the legacy `/tmp/*.sh` bypass pattern.
 
-1. Write the artifact body to `/tmp/harness-eval-$(date +%H%M%S).md` first (Write tool target = `/tmp`, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/sessions/$(date +%Y-%m-%d)/` (Bash target = `/tmp`, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
-
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling, #1016, #1045.
+Reference: R006 Sensitive Path Handling, R010 Universal bypassPermissions, #1101.
 
 
 ## 4-Metric Quantitative Layer (added v0.113.0, #1025)

package/templates/.claude/skills/harness-synthesizer/SKILL.md

@@ -96,15 +96,11 @@ harness:
 
 ### Tool: Writing artifacts under .claude/outputs/
 
-CC sensitive-path check inspects tool target paths and triggers permission prompts on `.claude/` regardless of `bypassPermissions` and allow rules (refs: #960, #961, #978, #981, #1016).
+Under `mode: "bypassPermissions"`, direct Write/Edit/Bash on `.claude/**` paths (including `.claude/outputs/sessions/`) is permitted (CC v2.1.121+, #1101) — no `/tmp` wrapping is needed.
 
-To write harness-synthesizer results under `.claude/outputs/sessions/`:
+Write harness-synthesizer results directly to `.claude/outputs/sessions/$(date +%Y-%m-%d)/harness-synthesizer-$(date +%H%M%S).md`. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification. Catastrophic shell operations (e.g., `rm -rf /`) remain blocked by independent safety guards. For CC < v2.1.121, see git history for the legacy `/tmp/*.sh` bypass pattern.
 
-1. Write the artifact body to `/tmp/harness-synthesizer-$(date +%H%M%S).md` first (Write tool target = `/tmp`, no sensitive-path trigger)
-2. Use a `/tmp/*.sh` Bash script to move/copy the file under `.claude/outputs/sessions/$(date +%Y-%m-%d)/` (Bash target = `/tmp`, script-internal `cp` to `.claude/` is not audited)
-3. Read-only Bash on `.claude/outputs/` (e.g., `cat`, `head`, `wc`) is allowed for verification
-
-Reference: `feedback_sensitive_path_tmp_bypass.md`, R006 sensitive-path handling, #1016, #1045.
+Reference: R006 Sensitive Path Handling, R010 Universal bypassPermissions, #1101.
 
 3. **Synthesize validation harness** — generate YAML harness matching agent's declared capabilities
 4. **Refine via evaluator-optimizer loop** — iterate harness against edge cases (3 rounds max)

package/templates/.claude/skills/help/SKILL.md

@@ -73,9 +73,6 @@ help --agents
 
 Available Agents:
 
-Orchestrator:
-  secretary        Manages manager agents
-
 Manager:
   mgr-creator          Creates new agents
   mgr-updater          Updates external sources and docs
@@ -98,7 +95,7 @@ Infra Engineer:
   infra-docker-expert       Docker containerization
   infra-aws-expert          AWS architecture
 
-Total: 15 agents
+Total: 49 agents
 ```
 
 ### Rules List
@@ -112,16 +109,16 @@ MUST (Never violate):
   R001  Safety Rules         Prohibited actions
   R002  Permission Rules     Tool tiers, file access
   R006  Agent Design         Structure, separation
+  R007  Agent Identification Display agent in responses
+  R008  Tool Identification  Display agent when using tools
+  R009  Parallel Execution   Max 4 parallel instances
 
 SHOULD (Strongly recommended):
   R003  Interaction Rules    Response format
   R004  Error Handling       Error levels, recovery
-  R007  Agent Identification Display agent in responses
-  R008  Tool Identification  Display agent when using tools
 
 MAY (Optional):
   R005  Optimization         Efficiency guidelines
-  R009  Parallel Execution   Max 4 parallel instances
 
-Total: 10 rules
+Total: 23 rules
 ```