@yeongjaeyou/claude-code-config 0.18.1 → 0.18.3

package/.claude/commands/ask-codex.md

@@ -75,6 +75,11 @@ You are a code review expert reviewing work in progress.
 ### CLAUDE.md
 [Project guidelines]
 
+### Work Guidelines
+Read and follow: .claude/guidelines/work-guidelines.md
+If .claude/rules/ exists, also read relevant rule files for module-specific guidance.
+(Use your file access to read these files directly)
+
 ### Directory Structure
 [tree output]
 
@@ -105,6 +110,14 @@ You are a code review expert reviewing work in progress.
 
 ### Summary
 [1-2 sentence conclusion]
+
+### Beyond the Question (Evidence-Based)
+If you identify improvements beyond the direct question:
+- Alternative approaches FOUND IN this codebase (with file:line)
+- Architectural patterns ALREADY USED that could apply
+- Potential optimizations BASED ON actual code analysis
+
+DO NOT suggest generic best practices without codebase evidence.
 ```
 
 ### 2. Parse Response

package/.claude/commands/ask-gemini.md

@@ -79,6 +79,11 @@ You are a code review expert reviewing work in progress.
 ### CLAUDE.md
 [Project guidelines]
 
+### Work Guidelines
+Read and follow: .claude/guidelines/work-guidelines.md
+If .claude/rules/ exists, also read relevant rule files for module-specific guidance.
+(Use your file access to read these files directly)
+
 ### Directory Structure
 [tree output]
 
@@ -109,6 +114,14 @@ You are a code review expert reviewing work in progress.
 
 ### Summary
 [1-2 sentence conclusion]
+
+### Beyond the Question (Evidence-Based)
+If you identify improvements beyond the direct question:
+- Alternative approaches FOUND IN this codebase (with file:line)
+- Architectural patterns ALREADY USED that could apply
+- Potential optimizations BASED ON actual code analysis
+
+DO NOT suggest generic best practices without codebase evidence.
 ```
 
 ### 2. Parse Response

package/.claude/commands/council.md

@@ -92,6 +92,10 @@ council_member:
         context: "why this is relevant"
     caveats:  # optional
       - "potential limitation or edge case"
+    beyond_question:  # optional, evidence-based only
+      - insight: "improvement opportunity"
+        evidence: "file:line or codebase reference"
+        rationale: "why this is relevant to the question context"
   # Round 2+ additional fields
   gaps:
     - "aspect not fully addressed"
@@ -103,6 +107,14 @@ council_member:
 - Sub-agents that fail to follow this schema will have their results flagged
 - Missing required fields trigger re-query in next round
 
+**Beyond the Question (Evidence-Based Only):**
+Council members may suggest improvements beyond the direct question, but ONLY with:
+- Specific file:line references from the codebase
+- Evidence from actual code analysis
+- Clear connection to the question context
+
+Generic best practices without codebase evidence are NOT accepted.
+
 ---
 
 ## Context Gathering (Before Round 1)
@@ -118,6 +130,8 @@ Model-specific guidelines (project root):
 - ./CLAUDE.md (Claude Opus/Sonnet)
 - ./AGENTS.md (Codex)
 - ./gemini.md (Gemini)
+- .claude/guidelines/work-guidelines.md (All models - style and response guidelines)
+- .claude/rules/*.md (All models - module-specific rules, if directory exists)
 ```
 
 **Conditional Code Exploration:**
@@ -235,6 +249,7 @@ Task(model="opus", subagent_type="general-purpose", run_in_background: true):
 
     ## Guidelines
     Read and follow ./CLAUDE.md project guidelines.
+    Read and follow: .claude/guidelines/work-guidelines.md for style guidelines.
     You have access to MCP tools. Use them actively to gather accurate information.
 
     ## Question
@@ -257,7 +272,8 @@ Task(model="opus", subagent_type="general-purpose", run_in_background: true):
 **Claude Sonnet:**
 ```
 Task(model="sonnet", subagent_type="general-purpose", run_in_background: true):
-  prompt: [Same structure as Opus]
+  prompt: [Same structure as Opus, including:
+    Read and follow: .claude/guidelines/work-guidelines.md for style guidelines.]
 ```
 
 **Codex:**
@@ -275,6 +291,7 @@ Task(subagent_type="general-purpose", run_in_background: true):
 
     ## Guidelines
     Read and follow ./AGENTS.md project guidelines.
+    Read and follow: .claude/guidelines/work-guidelines.md for style guidelines.
     You have access to MCP tools. Use them actively to gather accurate information.
 
     ## Question
@@ -308,6 +325,7 @@ Task(subagent_type="general-purpose", run_in_background: true):
 
     ## Guidelines
     Read and follow ./gemini.md project guidelines.
+    Read and follow: .claude/guidelines/work-guidelines.md for style guidelines.
     You have access to MCP tools. Use them actively to gather accurate information.
 
     ## Question

package/.claude/commands/gh/auto-review-loop.md

@@ -17,138 +17,163 @@ Automatically fix and push until CodeRabbit review comments reach zero.
 | Limit | Value | Purpose |
 |-------|-------|---------|
 | MAX_ITERATIONS | 10 | Prevent infinite loops |
-| POLL_TIMEOUT | 300s | Wait for CodeRabbit analysis |
+| POLL_TIMEOUT | 300s | Wait for checks to complete |
 | POLL_INTERVAL | 30s | Status check frequency |
 
-## Workflow (2-API Design)
+## Workflow
 
 ```
-┌─────────────────────────────────────────────────────────────┐
-│  1. Init → 2. Wait (Status API) → 3. Check (GraphQL) → 4. Fix │
-│                 ↑                                        ↓    │
-│                 └──────────── push & loop ───────────────┘    │
-└─────────────────────────────────────────────────────────────┘
+┌───────────────────────────────────────────────────────────────┐
+│  1. Init                                                      │
+│       ↓                                                       │
+│  ┌─────────────────────────────────────────────────────────┐  │
+│  │  2. Wait for All Checks (gh pr checks)                  │  │
+│  │       ↓                                                 │  │
+│  │  3. Check Unresolved Threads (GraphQL)                  │  │
+│  │       ↓                                                 │  │
+│  │  4. Fix, Commit, Push                                   │  │
+│  │       ↓                                                 │  │
+│  │  [loop back to 2 until resolved or max iterations]      │  │
+│  └─────────────────────────────────────────────────────────┘  │
+│       ↓                                                       │
+│  5. Output result                                             │
+└───────────────────────────────────────────────────────────────┘
 ```
 
 ### Step 1: Initialize
 
 ```bash
-# Get PR info
 PR_NUMBER=${ARGUMENTS:-$(gh pr view --json number -q .number 2>/dev/null)}
 REPO=$(gh repo view --json nameWithOwner -q .nameWithOwner)
-SHA=$(gh pr view $PR_NUMBER --json commits --jq '.commits[-1].oid')
-
 ITERATION=1
+MAX_ITERATIONS=10
 ```
 
-### Step 2: Wait for CodeRabbit (Status API)
+### Step 2: Wait for All Checks
 
-CodeRabbit uses Commit Status API, not Check Runs.
+`gh pr checks`로 CodeRabbit과 Actions 상태를 한 번에 확인.
 
 ```bash
-wait_for_coderabbit() {
-  local sha=$1 elapsed=0
+wait_for_all_checks() {
+  local elapsed=0
 
   while [ $elapsed -lt 300 ]; do
-    STATE=$(gh api repos/$REPO/commits/$sha/status \
-      --jq '.statuses[] | select(.context=="CodeRabbit") | .state' 2>/dev/null)
-
-    case "$STATE" in
-      success) return 0 ;;  # Review complete
-      pending) ;;           # Still analyzing
-      *)                    # No status yet or error
-        # Check if CodeRabbit review exists at all
-        HAS_REVIEW=$(gh pr view $PR_NUMBER --json reviews \
-          --jq '[.reviews[] | select(.author.login | contains("coderabbit"))] | length')
-        [ "$HAS_REVIEW" -gt 0 ] && return 0
-        ;;
-    esac
-
-    sleep 30
-    elapsed=$((elapsed + 30))
+    CHECKS=$(gh pr checks $PR_NUMBER 2>/dev/null)
+
+    # pending/in_progress 있으면 대기
+    if echo "$CHECKS" | grep -qiE "pending|in_progress|running"; then
+      sleep 30
+      elapsed=$((elapsed + 30))
+      continue
+    fi
+
+    # fail 있으면 에러 (CodeRabbit fail은 제외 - 리뷰 코멘트가 있다는 의미)
+    if echo "$CHECKS" | grep -v "CodeRabbit" | grep -qi "fail"; then
+      return 2  # CI failed
+    fi
+
+    return 0  # All checks completed
   done
 
   return 1  # Timeout
 }
-
-wait_for_coderabbit "$SHA" || {
-  echo "REVIEW_SKIPPED: CodeRabbit not responding within timeout"
-  exit 0
-}
 ```
 
 ### Step 3: Check Unresolved Threads (GraphQL)
 
-Single query for count + content. This is the ground truth.
+CodeRabbit 리뷰 코멘트 중 미해결 건수 확인. This is the ground truth.
 
 ```bash
-GRAPHQL_QUERY='
-query($owner: String!, $repo: String!, $pr: Int!) {
-  repository(owner: $owner, name: $repo) {
-    pullRequest(number: $pr) {
-      reviewThreads(first: 100) {
-        nodes {
-          isResolved
-          path
-          line
-          comments(first: 1) {
-            nodes {
-              author { login }
-              body
+check_unresolved_threads() {
+  RESULT=$(gh api graphql -f query='query {
+    repository(owner: "'"${REPO%/*}"'", name: "'"${REPO#*/}"'") {
+      pullRequest(number: '"$PR_NUMBER"') {
+        reviewThreads(first: 100) {
+          nodes {
+            isResolved
+            path
+            line
+            comments(first: 1) {
+              nodes {
+                author { login }
+                body
+              }
             }
           }
         }
       }
     }
-  }
-}'
+  }')
 
-# Filter: unresolved + coderabbit + not nitpick
-JQ_FILTER='[.data.repository.pullRequest.reviewThreads.nodes[] |
-  select(.isResolved == false) |
-  select(.comments.nodes[0].author.login | contains("coderabbit")) |
-  select(.comments.nodes[0].body | (contains("<!-- nitpick -->") or contains("[nitpick]")) | not)]'
+  # Filter: unresolved + coderabbit + not nitpick
+  UNRESOLVED=$(echo "$RESULT" | jq '[.data.repository.pullRequest.reviewThreads.nodes[] |
+    select(.isResolved == false) |
+    select(.comments.nodes[0].author.login | ascii_downcase | contains("coderabbit")) |
+    select(.comments.nodes[0].body | (contains("[nitpick]") or contains("nitpick")) | not)]')
 
-RESULT=$(gh api graphql -f query="$GRAPHQL_QUERY" \
-  -f owner="${REPO%/*}" -f repo="${REPO#*/}" -F pr="$PR_NUMBER")
+  UNRESOLVED_COUNT=$(echo "$UNRESOLVED" | jq 'length')
 
-UNRESOLVED=$(echo "$RESULT" | jq "$JQ_FILTER | length")
+  # Extract comments for fixing
+  COMMENTS=$(echo "$UNRESOLVED" | jq -r '.[] |
+    "### File: \(.path):\(.line)\n\n\(.comments.nodes[0].body)\n\n---\n"')
+}
+```
 
-if [ "$UNRESOLVED" -eq 0 ]; then
-  echo "REVIEW_COMPLETE: all comments resolved"
-  exit 0
-fi
+### Step 4: Fix, Commit, Push
 
-# Extract comments for fixing
-COMMENTS=$(echo "$RESULT" | jq -r "$JQ_FILTER | .[] |
-  \"### File: \(.path):\(.line)\n\n\(.comments.nodes[0].body)\n\n---\n\"")
+```bash
+fix_and_push() {
+  # For each comment: analyze and fix
+  # Follow code-review.md guidelines for auto-fix vs manual
+
+  # After fixes applied:
+  if [ -z "$(git status --porcelain)" ]; then
+    return 1  # No changes made
+  fi
+
+  git add -u
+  git commit -m "fix: address CodeRabbit review (iteration $ITERATION)"
+  git push
+}
 ```
 
-### Step 4: Fix, Commit, Loop
+### Main Loop
 
 ```bash
-# For each comment: analyze and fix
-# Follow code-review.md guidelines for auto-fix vs manual
-
-# After fixes applied:
-if [ -z "$(git status --porcelain)" ]; then
-  echo "REVIEW_COMPLETE: no changes needed"
-  exit 0
-fi
-
-git add -u
-git commit -m "fix: address CodeRabbit review (iteration $ITERATION)"
-git push
-
-ITERATION=$((ITERATION + 1))
-[ $ITERATION -gt 10 ] && {
-  echo "REVIEW_INCOMPLETE: max iterations (10) reached, $UNRESOLVED comments remaining"
-  exit 1
-}
+while [ $ITERATION -le $MAX_ITERATIONS ]; do
+  # Step 2: Wait for all checks
+  wait_for_all_checks
+  WAIT_RESULT=$?
+
+  if [ $WAIT_RESULT -eq 1 ]; then
+    echo "REVIEW_SKIPPED: checks not completing within timeout"
+    exit 0
+  elif [ $WAIT_RESULT -eq 2 ]; then
+    echo "REVIEW_BLOCKED: CI checks failed"
+    exit 1
+  fi
+
+  # Step 3: Check unresolved threads
+  check_unresolved_threads
+
+  if [ "$UNRESOLVED_COUNT" -eq 0 ]; then
+    echo "REVIEW_COMPLETE: all comments resolved"
+    exit 0
+  fi
+
+  # Step 4: Fix and push
+  # ... fix logic here using COMMENTS ...
+
+  fix_and_push || {
+    echo "REVIEW_COMPLETE: no changes needed"
+    exit 0
+  }
+
+  ITERATION=$((ITERATION + 1))
+done
 
-# Update SHA and loop back to Step 2
-SHA=$(git rev-parse HEAD)
-# Go to Step 2
+echo "REVIEW_INCOMPLETE: max iterations ($MAX_ITERATIONS) reached, $UNRESOLVED_COUNT comments remaining"
+exit 1
 ```
 
 ## Output Format
@@ -158,8 +183,10 @@ End with exactly one of:
 | Output | Meaning |
 |--------|---------|
 | `REVIEW_COMPLETE: all comments resolved` | Success |
+| `REVIEW_COMPLETE: no changes needed` | No fixes required |
 | `REVIEW_INCOMPLETE: max iterations reached, N comments remaining` | Hit limit |
-| `REVIEW_SKIPPED: CodeRabbit not responding within timeout` | No review |
+| `REVIEW_SKIPPED: checks not completing within timeout` | Timeout |
+| `REVIEW_BLOCKED: CI checks failed` | CI failure |
 
 ## Fix Guidelines
 
@@ -175,4 +202,4 @@ Read `.claude/commands/code-review.md` for:
 | PR not found | Exit with error |
 | API rate limit | Increase interval, retry |
 | Git push failed | Report and stop |
-| GraphQL error | Fall back to pr view |
+| CI checks failed | Exit with REVIEW_BLOCKED |

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yeongjaeyou/claude-code-config",
-  "version": "0.18.1",
+  "version": "0.18.3",
   "description": "Claude Code CLI custom commands, agents, and skills",
   "bin": {
     "claude-code-config": "./bin/cli.js"