oh-my-customcode 0.51.2 → 0.53.0

package/README.md

@@ -13,7 +13,7 @@
 
 **[한국어 문서 (Korean)](./README_ko.md)**
 
-45 agents. 91 skills. 21 rules. One command.
+45 agents. 92 skills. 21 rules. One command.
 
 ```bash
 npm install -g oh-my-customcode && cd your-project && omcustom init
@@ -52,6 +52,10 @@ oh-my-customcode is built on two ideas:
 
 Skills are source. Agents are compiled output. Sauron verifies the build. This separation means skills evolve independently of agents, and agents can be recompiled from updated skills at any time.
 
+<p align="center">
+  <img src="assets/diagrams/05-compilation-metaphor.png" alt="Compilation Metaphor" width="700" />
+</p>
+
 **2. If it can't be done, make it work.**
 
 When no specialist exists for a task, oh-my-customcode does not fail. It creates one.
@@ -85,6 +89,10 @@ User (natural language)
 
 Four routing skills cover the full domain:
 
+<p align="center">
+  <img src="assets/diagrams/01-system-architecture.png" alt="System Architecture" width="700" />
+</p>
+
 | Routing Skill | Routes To |
 |--------------|-----------|
 | secretary-routing | Manager agents (mgr-*), system agents (sys-*) |
@@ -138,7 +146,7 @@ Each agent declares its tools, model, memory scope, and limitations in YAML fron
 
 ---
 
-### Skills (91)
+### Skills (92)
 
 | Category | Count | Includes |
 |----------|-------|----------|

package/dist/cli/index.js

@@ -9323,7 +9323,7 @@ var init_package = __esm(() => {
   package_default = {
     name: "oh-my-customcode",
     workspaces: ["packages/*"],
-    version: "0.51.2",
+    version: "0.53.0",
     description: "Batteries-included agent harness for Claude Code",
     type: "module",
     bin: {

package/dist/index.js

@@ -1668,7 +1668,7 @@ import { join as join6 } from "node:path";
 var package_default = {
   name: "oh-my-customcode",
   workspaces: ["packages/*"],
-  version: "0.51.2",
+  version: "0.53.0",
   description: "Batteries-included agent harness for Claude Code",
   type: "module",
   bin: {

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "oh-my-customcode",
   "workspaces": ["packages/*"],
-  "version": "0.51.2",
+  "version": "0.53.0",
   "description": "Batteries-included agent harness for Claude Code",
   "type": "module",
   "bin": {

package/templates/.claude/hooks/hooks.json

@@ -324,6 +324,16 @@
         ],
         "description": "Batch-save agent outcomes to eval-core DB on session end (advisory, exit 0)"
       },
+      {
+        "matcher": "*",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "bash .claude/hooks/scripts/feedback-collector.sh"
+          }
+        ],
+        "description": "Auto-extract failure patterns from session outcomes (advisory, exit 0)"
+      },
       {
         "matcher": "*",
         "hooks": [

package/templates/.claude/hooks/scripts/feedback-collector.sh

@@ -0,0 +1,74 @@
+#!/usr/bin/env bash
+# feedback-collector.sh — Auto-extract failure patterns from session outcomes
+# Advisory-only: always exits 0, never blocks session end
+
+set -euo pipefail
+
+# Pass through stdin (Stop hook protocol)
+cat > /dev/null
+
+# Dependencies check
+command -v jq >/dev/null 2>&1 || exit 0
+command -v sqlite3 >/dev/null 2>&1 || exit 0
+
+# PID scoping
+OUTCOMES_FILE="/tmp/.claude-task-outcomes-${PPID}"
+[ -f "$OUTCOMES_FILE" ] || exit 0
+
+# DB path
+DB_PATH="${HOME}/.config/oh-my-customcode/eval-core.sqlite"
+[ -f "$DB_PATH" ] || exit 0
+
+# Count failures per agent type
+declare -A FAILURE_COUNTS
+declare -A TOTAL_COUNTS
+
+while IFS= read -r line; do
+  agent_type=$(echo "$line" | jq -r '.agent_type // empty' 2>/dev/null) || continue
+  outcome=$(echo "$line" | jq -r '.outcome // empty' 2>/dev/null) || continue
+  [ -z "$agent_type" ] && continue
+
+  TOTAL_COUNTS[$agent_type]=$(( ${TOTAL_COUNTS[$agent_type]:-0} + 1 ))
+  if [ "$outcome" = "failure" ]; then
+    FAILURE_COUNTS[$agent_type]=$(( ${FAILURE_COUNTS[$agent_type]:-0} + 1 ))
+  fi
+done < "$OUTCOMES_FILE"
+
+# Detect repeated failure agents (3+ failures)
+TIMESTAMP=$(date -u +"%Y-%m-%dT%H:%M:%SZ")
+INSERTED=0
+
+for agent_type in "${!FAILURE_COUNTS[@]}"; do
+  count=${FAILURE_COUNTS[$agent_type]}
+  total=${TOTAL_COUNTS[$agent_type]:-0}
+  [ "$count" -lt 3 ] && continue
+
+  # Determine confidence
+  if [ "$count" -ge 5 ]; then
+    confidence="high"
+  elif [ "$count" -ge 3 ]; then
+    confidence="medium"
+  else
+    confidence="low"
+  fi
+
+  # Determine action type
+  if [ "$count" -ge 5 ]; then
+    action_type="escalate"
+  else
+    action_type="augment"
+  fi
+
+  failure_rate=$(awk "BEGIN {printf \"%.2f\", $count/$total}")
+  description="Agent '${agent_type}' failed ${count}/${total} times (${failure_rate} failure rate) in session"
+
+  sqlite3 "$DB_PATH" "INSERT INTO improvementActions (targetType, targetName, actionType, description, confidence, feedbackSource, status, createdAt) VALUES ('agent', '${agent_type}', '${action_type}', '${description}', '${confidence}', 'outcome_derived', 'proposed', '${TIMESTAMP}');" 2>/dev/null || true
+
+  INSERTED=$((INSERTED + 1))
+done
+
+if [ "$INSERTED" -gt 0 ]; then
+  echo "[feedback-collector] Extracted ${INSERTED} failure pattern(s) from session outcomes" >&2
+fi
+
+exit 0

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

@@ -135,6 +135,12 @@ The skill's steps are followed, but agent spawning uses Teams when criteria are
      Agent(researcher-1) → Analysis 1  ┐
      Agent(researcher-2) → Analysis 2  ├─ ALL spawned together
      Agent(researcher-3) → Analysis 3  ┘
+
+❌ WRONG: Completed member modifies other member's files
+   svelte-projects completes task → browses TaskList → edits agent-teams-advisor.sh (hook-fixer's scope)
+
+✓ CORRECT: Completed member reports and waits
+   svelte-projects completes task → SendMessage("Task complete") → waits silently
 ```
 
 ## Cost Guidelines
@@ -205,12 +211,14 @@ When a team member is blocked by task dependencies:
 | Silent wait | Agent already spawned, short wait expected | Minimal overhead |
 | Reassign | Agent blocked >2 min with no progress | Reuse agent for unblocked work |
 
-### Prompt Guidelines for Blocked Agents
+### Prompt Guidelines for Team Members
 
 When spawning agents that may be blocked:
 1. Include explicit instruction: "If your task is blocked, wait silently. Do NOT send periodic status messages."
 2. Set check interval: "Check TaskList once per minute, not continuously."
 3. Prefer deferred spawn when the dependency resolution time is unpredictable.
+4. Post-completion instruction: "After completing your task, report via SendMessage and wait. Do NOT explore or modify files outside your scope."
+5. Explicit scope boundary: "Your scope is limited to: {file list or directory}. Do NOT modify files outside this scope."
 
 ### Anti-Pattern: Idle Polling
 
@@ -225,6 +233,36 @@ When spawning agents that may be blocked:
    docker-dev spawned with: "Wait silently if blocked. Check TaskList once per minute."
 ```
 
+## Post-Completion Scope Constraint
+
+When a team member completes its assigned task:
+
+| Behavior | Correct | Wrong |
+|----------|---------|-------|
+| Task completed | Report completion via SendMessage, wait silently | Browse TaskList for other work |
+| No more tasks | Exit or wait for team shutdown | Explore/modify files outside original scope |
+| See unfinished work | Report to team lead, do NOT self-assign | Edit files that belong to other members |
+
+### Self-Check (After Task Completion)
+
+```
+╔══════════════════════════════════════════════════════════════════╗
+║  AFTER COMPLETING YOUR ASSIGNED TASK:                            ║
+║                                                                   ║
+║  1. Did I complete ONLY my assigned task?                        ║
+║     YES → Report completion                                      ║
+║     NO  → Revert scope-violation changes                         ║
+║                                                                   ║
+║  2. Are there files modified outside my task scope?              ║
+║     YES → This is a violation — revert                           ║
+║     NO  → Good                                                    ║
+║                                                                   ║
+║  3. Am I about to explore/modify files for "other tasks"?        ║
+║     YES → STOP. Report to team lead instead                      ║
+║     NO  → Good. Wait silently or exit                            ║
+╚══════════════════════════════════════════════════════════════════╝
+```
+
 ## Lifecycle
 
 ```

package/templates/.claude/skills/omcustom-improve-report/SKILL.md

@@ -0,0 +1,65 @@
+---
+name: omcustom-improve-report
+description: Read-only report of improvement suggestions from eval-core analysis engine
+scope: harness
+user-invocable: true
+---
+
+# Improve Report
+
+Display improvement suggestions from eval-core analysis engine as read-only report. No file modifications, no GitHub mutations.
+
+## Purpose
+
+Surface actionable improvement suggestions gathered by the eval-core analysis engine. Reads from eval-core's local database and renders insights as structured markdown. Useful for understanding routing quality, skill effectiveness, and agent usage patterns.
+
+## Usage
+
+```
+/omcustom:improve-report
+```
+
+## Workflow
+
+### Step 1: Check eval-core availability
+
+```bash
+command -v eval-core 2>/dev/null || ls node_modules/.bin/eval-core 2>/dev/null
+```
+
+If not found → skip to Step 4 (fallback).
+
+### Step 2: Run analysis
+
+```bash
+eval-core analyze --format markdown
+```
+
+### Step 3: Data sufficiency check
+
+Count sessions recorded from the output metadata:
+
+| Sessions recorded | Confidence | Behavior |
+|-------------------|------------|----------|
+| < 5 | `[confidence: low]` | Show results with message: "최소 5세션 이상의 데이터가 필요합니다. 현재 데이터로도 초기 분석은 가능하지만, 정확도는 제한적입니다." |
+| 5–20 | `[confidence: medium]` | Show results with medium confidence note |
+| > 20 | `[confidence: high]` | Show results normally |
+
+### Step 4: Display
+
+If output contains data → display as structured markdown with confidence annotation prepended.
+
+If command not found OR output is empty:
+
+```
+[omcustom:improve-report] 데이터 없음
+eval-core가 설치되어 있지 않거나 아직 충분한 세션 데이터가 수집되지 않았습니다.
+eval-core를 설치하거나 더 많은 세션을 진행한 후 다시 실행하세요.
+```
+
+## Notes
+
+- **Read-only**: No file modifications, no GitHub issue creation, no external mutations
+- **R010 compliant**: Only executes Bash for read-only analysis — no write operations delegated
+- **Graceful fallback**: Falls back silently if eval-core or its DB doesn't exist
+- **Confidence-aware**: Reports data sufficiency alongside results per R011 conventions

package/templates/CLAUDE.md

@@ -101,6 +101,7 @@ oh-my-customcode로 구동됩니다.
 | `/omcustom:update-external` | 외부 소스에서 에이전트 업데이트 |
 | `/omcustom:audit-agents` | 에이전트 의존성 감사 |
 | `/omcustom:fix-refs` | 깨진 참조 수정 |
+| `/omcustom:improve-report` | eval-core 기반 개선 현황 리포트 |
 | `/omcustom-takeover` | 기존 에이전트/스킬에서 canonical spec 추출 |
 | `/adversarial-review` | 공격자 관점 보안 코드 리뷰 |
 | `/ambiguity-gate` | 요청 모호성 분석 및 명확화 질문 (ouroboros 패턴) |

package/templates/manifest.json

@@ -1,5 +1,5 @@
 {
-  "version": "0.51.2",
+  "version": "0.53.0",
   "lastUpdated": "2026-03-16T00:00:00.000Z",
   "components": [
     {
@@ -18,7 +18,7 @@
       "name": "skills",
       "path": ".claude/skills",
       "description": "Reusable skill modules (includes slash commands)",
-      "files": 91
+      "files": 92
     },
     {
       "name": "guides",