all-for-claudecode 2.10.0 → 2.12.0

package/.claude-plugin/marketplace.json

@@ -6,14 +6,14 @@
   },
   "metadata": {
     "description": "Automated pipeline for Claude Code — spec → plan → implement → review → clean",
-    "version": "2.10.0"
+    "version": "2.12.0"
   },
   "plugins": [
     {
       "name": "afc",
       "source": "./",
       "description": "Automated pipeline for Claude Code. Automates the full development cycle: spec → plan → implement → review → clean.",
-      "version": "2.10.0",
+      "version": "2.12.0",
       "category": "automation",
       "tags": ["pipeline", "automation", "spec", "plan", "implement", "review", "critic-loop"]
     }

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "afc",
-  "version": "2.10.0",
+  "version": "2.12.0",
   "description": "Automated pipeline for Claude Code. Automates the full development cycle: spec → plan → implement → review → clean.",
   "author": { "name": "jhlee0409", "email": "relee6203@gmail.com" },
   "homepage": "https://github.com/jhlee0409/all-for-claudecode",

package/MIGRATION.md

@@ -20,7 +20,7 @@
 - If you previously used `/afc:analyze` for artifact consistency checking, use `/afc:validate` instead (or let the pipeline invoke it automatically).
 - Agent memory files (`.claude/agent-memory/afc-architect/MEMORY.md`, `.claude/agent-memory/afc-security/MEMORY.md`) exceeding 100 lines will be auto-pruned on next pipeline run.
 - Memory subdirectories exceeding rotation thresholds will be auto-pruned during the Clean phase.
-- Run `/afc:init` to update the `AFC:VERSION` tag in your CLAUDE.md.
+- Run `/afc:setup` to update the `AFC:VERSION` tag in your CLAUDE.md.
 
 ---
 
@@ -77,7 +77,7 @@ Replace the `SELFISH:START` / `SELFISH:END` block in your project's CLAUDE.md:
 - Change all `selfish:` command references → `afc:`
 - Change `selfish-architect` / `selfish-security` → `afc-architect` / `afc-security`
 
-Or simply run `/afc:init` to regenerate the block.
+Or simply run `/afc:setup` to regenerate the block.
 
 #### 4. Migrate state files (if pipeline was active)
 

package/README.md

@@ -28,7 +28,8 @@ npx all-for-claudecode
 Then:
 
 ```
-/afc:init                              # Detect your stack, generate config
+/afc:setup                             # Set up global routing in ~/.claude/CLAUDE.md
+/afc:init                              # Detect your stack, generate project config
 /afc:auto "Add user authentication"    # Run the full pipeline
 ```
 
@@ -52,6 +53,10 @@ Spec (1/5) → Plan (2/5) → Implement (3/5) → Review (4/5) → Clean (5/5)
   │              └─ File change map, ADR recording, research persistence
   └─ Acceptance criteria, pre-implementation gates
 
+Skill Advisor checkpoints (A–E) between phases dynamically invoke
+auxiliary skills (ideate, consult, architect, security, test, qa, learner)
+when LLM evaluation detects they would add value. Budget: max 5 per run.
+
 Hooks run automatically at each step.
 CI failure → debug-based RCA (not blind retry).
 Critic Loops verify quality at each gate until convergence.
@@ -111,7 +116,8 @@ Performance: ✓ no N+1 queries
 | `/afc:clean` | Pipeline artifact cleanup and codebase hygiene |
 | `/afc:research` | Technical research with persistent storage |
 | `/afc:debug` | Bug diagnosis and fix |
-| `/afc:init` | Project setup — detects stack and generates config |
+| `/afc:init` | Project setup — detects stack and generates project config |
+| `/afc:setup` | Global CLAUDE.md configuration — injects/updates AFC routing block |
 | `/afc:doctor` | Diagnose project health and plugin setup |
 | `/afc:architect` | Architecture analysis (persistent memory) |
 | `/afc:security` | Security scan (persistent memory, isolated worktree) |
@@ -244,10 +250,12 @@ Dependencies are tracked via DAG. CI gate + Mini-Review + Auto-Checkpoint run at
 ## Configuration
 
 ```
-/afc:init
+/afc:setup    # One-time: inject routing block into ~/.claude/CLAUDE.md
+/afc:init     # Per-project: detect stack and generate .claude/afc.config.md
 ```
 
-Auto-detects your tech stack (package manager, framework, architecture, testing, linting) and generates `.claude/afc.config.md` with CI commands, architecture rules, and code style conventions. No manual preset selection needed — the init command analyzes your project structure directly.
+- **`/afc:setup`** — Manages the AFC routing block in your global `~/.claude/CLAUDE.md`. Run once after install and again after plugin updates. Idempotent — skips if version matches.
+- **`/afc:init`** — Auto-detects your tech stack (package manager, framework, architecture, testing, linting) and generates project-local config files. No manual preset selection needed.
 
 ## FAQ
 

package/bin/cli.mjs

@@ -97,6 +97,7 @@ async function main() {
     console.log("  ✓ Installation complete!");
     console.log();
     console.log("  Next steps:");
+    console.log("    /afc:setup                   Set up global routing");
     console.log("    /afc:init                    Create project config");
     console.log('    /afc:auto "feature desc"      Run the pipeline');
     console.log();

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "all-for-claudecode",
-  "version": "2.10.0",
+  "version": "2.12.0",
   "description": "Claude Code plugin that automates the full dev cycle — spec, plan, implement, review, clean.",
   "bin": {
     "all-for-claudecode": "bin/cli.mjs"

package/scripts/afc-consistency-check.sh

@@ -113,6 +113,8 @@ check_agent_names() {
   local count=0
   local invalid=0
   for ref in $referenced_agents; do
+    # Skip dynamic template patterns (e.g., afc-{domain}-expert)
+    case "$ref" in *"{"*) continue ;; esac
     count=$((count + 1))
     if ! printf '%s\n' "$defined_agents" | grep -qxF "$ref"; then
       fail "subagent_type 'afc:$ref' referenced but no agents/$ref.md found"
@@ -262,7 +264,7 @@ check_command_docs() {
   [ -d "$skills_dir" ] || return
 
   local readme="$PROJECT_DIR/README.md"
-  local init_skill="$skills_dir/init/SKILL.md"
+  local setup_skill="$skills_dir/setup/SKILL.md"
   local claude_md="$PROJECT_DIR/CLAUDE.md"
   local issues=0
 
@@ -279,12 +281,12 @@ check_command_docs() {
       fi
     fi
 
-    # Sub-check B: init/SKILL.md should mention afc:{name} for user-invocable skills
+    # Sub-check B: setup/SKILL.md should mention afc:{name} for user-invocable skills
     local invocable
     invocable=$(get_cmd_field "$skill_file" "user-invocable")
-    if [ "$invocable" != "false" ] && [ -f "$init_skill" ]; then
-      if ! grep -qE "afc:${cmd_name}([^a-z0-9-]|$)" "$init_skill" 2>/dev/null; then
-        warn "Skill '$cmd_name' missing from init/SKILL.md skill routing"
+    if [ "$invocable" != "false" ] && [ -f "$setup_skill" ]; then
+      if ! grep -qE "afc:${cmd_name}([^a-z0-9-]|$)" "$setup_skill" 2>/dev/null; then
+        warn "Skill '$cmd_name' missing from setup/SKILL.md skill routing"
         issues=$((issues + 1))
       fi
     fi
@@ -301,7 +303,7 @@ check_command_docs() {
   done
 
   if [ "$issues" -eq 0 ]; then
-    ok "Skill docs: all skills referenced in README.md, init/SKILL.md, CLAUDE.md"
+    ok "Skill docs: all skills referenced in README.md, setup/SKILL.md, CLAUDE.md"
   fi
 }
 

package/scripts/afc-doctor.sh

@@ -55,6 +55,20 @@ section() {
   printf '\n%s\n' "$1"
 }
 
+# --- Header: show which plugin version is running ---
+if [ -f "$PLUGIN_ROOT/package.json" ]; then
+  if command -v jq >/dev/null 2>&1; then
+    RUNNING_VERSION=$(jq -r '.version // "unknown"' "$PLUGIN_ROOT/package.json" 2>/dev/null || echo "unknown")
+  else
+    RUNNING_VERSION=$(grep -o '"version"[[:space:]]*:[[:space:]]*"[^"]*"' "$PLUGIN_ROOT/package.json" 2>/dev/null | head -1 | sed 's/.*: *"//;s/"//' || echo "unknown")
+  fi
+  printf 'all-for-claudecode Doctor (v%s)\n' "$RUNNING_VERSION"
+  printf 'Plugin root: %s\n' "$PLUGIN_ROOT"
+else
+  printf 'all-for-claudecode Doctor (version unknown)\n'
+  printf 'Plugin root: %s\n' "$PLUGIN_ROOT"
+fi
+
 # --- Category 1: Environment ---
 section "Environment"
 
@@ -160,15 +174,15 @@ if [ -f "$GLOBAL_CLAUDE" ]; then
         if [ "$BLOCK_VERSION" = "$PLUGIN_VERSION" ]; then
           pass "Block version matches plugin ($PLUGIN_VERSION)"
         else
-          warn "all-for-claudecode block outdated (block: $BLOCK_VERSION, plugin: $PLUGIN_VERSION)" "run /afc:init to update"
+          warn "all-for-claudecode block outdated (block: $BLOCK_VERSION, plugin: $PLUGIN_VERSION)" "run /afc:setup to update"
         fi
       fi
     fi
   else
-    fail "all-for-claudecode block not found" "run /afc:init to inject all-for-claudecode block"
+    fail "all-for-claudecode block not found" "run /afc:setup to inject all-for-claudecode block"
   fi
 else
-  warn "No global ~/.claude/CLAUDE.md" "run /afc:init"
+  warn "No global ~/.claude/CLAUDE.md" "run /afc:setup"
 fi
 
 # --- Category 4: Legacy Migration ---
@@ -179,7 +193,7 @@ LEGACY_FOUND=false
 # Legacy CLAUDE.md block
 if [ -f "$GLOBAL_CLAUDE" ] && grep -q '<!-- SELFISH:START -->' "$GLOBAL_CLAUDE" 2>/dev/null; then
   LEGACY_FOUND=true
-  warn "Legacy SELFISH:START block in ~/.claude/CLAUDE.md" "run /afc:init (will replace)"
+  warn "Legacy SELFISH:START block in ~/.claude/CLAUDE.md" "run /afc:setup (will replace)"
 fi
 
 # Legacy config

package/scripts/session-start-context.sh

@@ -73,7 +73,7 @@ if [ -f "$PLUGIN_ROOT/package.json" ]; then
     if [ -f "$GLOBAL_CLAUDE" ]; then
       BLOCK_VERSION=$(grep -o 'AFC:VERSION:[0-9][0-9.]*' "$GLOBAL_CLAUDE" 2>/dev/null | head -1 | sed 's/AFC:VERSION://' || true)
       if [ -n "${BLOCK_VERSION:-}" ] && [ "$BLOCK_VERSION" != "$PLUGIN_VERSION" ]; then
-        OUTPUT="${OUTPUT:+$OUTPUT | }[AFC VERSION MISMATCH] v$PLUGIN_VERSION installed but CLAUDE.md block is v$BLOCK_VERSION. Run /afc:init to update."
+        OUTPUT="${OUTPUT:+$OUTPUT | }[AFC VERSION MISMATCH] v$PLUGIN_VERSION installed but CLAUDE.md block is v$BLOCK_VERSION. Run /afc:setup to update."
       fi
     fi
   fi

package/skills/analyze/SKILL.md

@@ -32,16 +32,18 @@ If neither rules file nor config exists: read `CLAUDE.md` for architecture info.
 
 ### 1. Parse Analysis Intent
 
-Classify `$ARGUMENTS` into one of these analysis modes:
+Read the user's question semantically. What does the user actually want to understand? Select the mode that best serves their learning goal.
 
-| Mode | Trigger Keywords | Focus |
-|------|-----------------|-------|
-| **Root Cause** | "why", "cause", "bug", "error", "broken" | Error trace → data flow → hypothesis |
-| **Structural** | "how", "architecture", "flow", "trace", "structure" | Component relationships, call graphs, data flow |
-| **Exploratory** | "what", "find", "where", "which", "list" | File/function discovery, pattern matching |
-| **Comparative** | "difference", "compare", "vs", "between" | Side-by-side analysis of implementations |
+| Mode | When to select | Focus |
+|------|---------------|-------|
+| **Root Cause** | User wants to understand WHY something is broken, failing, or behaving unexpectedly — the goal is diagnosing a problem | Error trace → data flow → hypothesis |
+| **Structural** | User wants to understand HOW a system is built or how components relate — the goal is comprehension of design or flow | Component relationships, call graphs, data flow |
+| **Exploratory** | User wants to discover WHAT exists in the codebase — the goal is finding, listing, or locating things | File/function discovery, pattern matching |
+| **Comparative** | User wants to understand the DIFFERENCE between two or more things — the goal is contrast and tradeoff evaluation | Side-by-side analysis of implementations |
 
-If the intent doesn't clearly match a mode, default to **Exploratory**.
+If the question spans multiple modes, select the PRIMARY mode that best matches the user's core learning goal, and note secondary aspects to incorporate during analysis.
+
+If the intent doesn't clearly match a single mode, default to **Exploratory**.
 
 ### 2. Codebase Exploration
 

package/skills/architect/SKILL.md

@@ -53,8 +53,8 @@ Analyze `$ARGUMENTS` to identify the task type:
 
 Use Agent Teams for wide analysis scope (3+ modules) with parallel exploration:
 ```
-Task("analyze features/timeline", subagent_type: Explore)
-Task("analyze widgets/timeline", subagent_type: Explore)
+Task("analyze features/timeline", subagent_type: "Explore")
+Task("analyze widgets/timeline", subagent_type: "Explore")
 ```
 
 **Cross-Module Import Chain Verification** (after Explore agents return):
@@ -99,9 +99,9 @@ Structure analysis results and **print to console**:
 
 ### 4. Critic Loop
 
-> **Always** read `${CLAUDE_PLUGIN_ROOT}/docs/critic-loop-rules.md` first and follow it.
+> **Always** read `${CLAUDE_SKILL_DIR}/../../docs/critic-loop-rules.md` first and follow it.
 
-Run the critic loop until convergence. Safety cap: 7 passes (higher than the standard 5 because architecture analysis involves broader exploration across modules and layers).
+Run the critic loop until convergence. Safety cap: 5 passes.
 
 | Criterion | Validation |
 |-----------|------------|