prizmkit 1.1.91 → 1.1.93

package/bundled/dev-pipeline/run-refactor.sh

@@ -8,14 +8,14 @@ set -euo pipefail
 # execute refactors from a .prizmkit/plans/refactor-list.json specification.
 #
 # Usage:
-#   ./run-refactor.sh run [.prizmkit/plans/refactor-list.json]          Run all refactors
+#   ./run-refactor.sh run [.prizmkit/plans/refactor-list.json] [options] Run all refactors
 #   ./run-refactor.sh run <refactor-id> [options]       Run a single refactor
 #   ./run-refactor.sh status [.prizmkit/plans/refactor-list.json]       Show pipeline status
 #   ./run-refactor.sh reset                             Clear all state
 #
 # Environment Variables:
-#   MAX_RETRIES           Max retries per refactor (default: 3)
-#   SESSION_TIMEOUT       Session timeout in seconds (default: 0 = no limit)
+#   MAX_RETRIES           Max code retries per refactor (default: 3)
+#   MAX_INFRA_RETRIES     Max infrastructure/provider retries per refactor (default: 3)
 #   AI_CLI                AI CLI command name (auto-detected: cbc, claude, or codex)
 #   CODEBUDDY_CLI         Legacy alias for AI_CLI (deprecated, use AI_CLI instead)
 #   PRIZMKIT_PLATFORM     Platform hint for custom AI_CLI wrappers: codebuddy, claude, or codex
@@ -23,9 +23,6 @@ set -euo pipefail
 #   HEARTBEAT_INTERVAL    Heartbeat log interval in seconds (default: 30)
 #   STALE_KILL_THRESHOLD   Auto-kill session after N seconds of no progress (default: 900)
 #   HEARTBEAT_STALE_THRESHOLD  Heartbeat stale threshold in seconds (default: 600)
-#   LOG_CLEANUP_ENABLED   Run periodic log cleanup (default: 1)
-#   LOG_RETENTION_DAYS    Delete logs older than N days (default: 14)
-#   LOG_MAX_TOTAL_MB      Keep total logs under N MB via oldest-first cleanup (default: 1024)
 #   DEV_BRANCH            Custom dev branch name (default: auto-generated refactor/pipeline-{run_id})
 #   AUTO_PUSH             Auto-push to remote after successful refactor (default: 0). Set to 1 to enable.
 #   STOP_ON_FAILURE       Stop pipeline after a task exhausts all retries (default: 0). Set to 1 to stop.
@@ -43,13 +40,10 @@ SCRIPTS_DIR="$SCRIPT_DIR/scripts"
 
 # Configuration
 MAX_RETRIES=${MAX_RETRIES:-3}
-SESSION_TIMEOUT=${SESSION_TIMEOUT:-0}
+MAX_INFRA_RETRIES=${MAX_INFRA_RETRIES:-3}
 HEARTBEAT_STALE_THRESHOLD=${HEARTBEAT_STALE_THRESHOLD:-600}
 HEARTBEAT_INTERVAL=${HEARTBEAT_INTERVAL:-30}
 STALE_KILL_THRESHOLD=${STALE_KILL_THRESHOLD:-900}
-LOG_CLEANUP_ENABLED=${LOG_CLEANUP_ENABLED:-1}
-LOG_RETENTION_DAYS=${LOG_RETENTION_DAYS:-14}
-LOG_MAX_TOTAL_MB=${LOG_MAX_TOTAL_MB:-1024}
 VERBOSE=${VERBOSE:-0}
 MODEL=${MODEL:-""}
 PRIZMKIT_EFFORT=${PRIZMKIT_EFFORT:-""}
@@ -92,8 +86,9 @@ spawn_and_wait_session() {
     local bootstrap_prompt="$4"
     local session_dir="$5"
     local max_retries="$6"
-    local item_model="${7:-}"
-    local base_branch="${8:-main}"
+    local max_infra_retries="$7"
+    local item_model="${8:-}"
+    local base_branch="${9:-main}"
 
     local session_log="$session_dir/logs/session.log"
     local progress_json="$session_dir/logs/progress.json"
@@ -116,13 +111,6 @@ spawn_and_wait_session() {
     start_progress_parser "$session_log" "$progress_json" "$SCRIPTS_DIR"
     local parser_pid="${_PARSER_PID:-}"
 
-    # Timeout watchdog
-    local watcher_pid=""
-    if [[ $SESSION_TIMEOUT -gt 0 ]]; then
-        ( sleep "$SESSION_TIMEOUT" && kill -TERM "$cli_pid" 2>/dev/null ) &
-        watcher_pid=$!
-    fi
-
     # Heartbeat monitor (with stale-kill protection)
     start_heartbeat "$cli_pid" "$session_log" "$progress_json" "$HEARTBEAT_INTERVAL" "$STALE_KILL_THRESHOLD"
     local heartbeat_pid="${_HEARTBEAT_PID:-}"
@@ -136,22 +124,32 @@ spawn_and_wait_session() {
     fi
 
     # Cleanup
-    [[ -n "$watcher_pid" ]] && kill "$watcher_pid" 2>/dev/null || true
     stop_heartbeat "$heartbeat_pid"
     stop_progress_parser "$parser_pid"
-    [[ -n "$watcher_pid" ]] && wait "$watcher_pid" 2>/dev/null || true
 
     # Recover session log from backup if truncated by subagent stdout redirection
     prizm_recover_session_log "$session_log"
 
-    [[ $exit_code -eq 143 ]] && exit_code=124
-
     # Check for stale-kill marker (heartbeat killed the process due to no progress)
     local stale_kill_marker="$session_dir/logs/stale-kill.json"
     local was_stale_killed=false
+    local stale_kill_reason=""
     if [[ -f "$stale_kill_marker" ]]; then
         was_stale_killed=true
-        log_warn "Session was stale-killed by heartbeat monitor (no progress for too long)"
+        stale_kill_reason=$(python3 - "$stale_kill_marker" <<'PY' 2>/dev/null || true
+import json, sys
+try:
+    with open(sys.argv[1], encoding="utf-8") as fh:
+        print(json.load(fh).get("reason") or "")
+except Exception:
+    pass
+PY
+)
+        if [[ "$stale_kill_reason" == "error_loop" ]]; then
+            log_warn "Session was killed by heartbeat monitor due to repeated read-offset/wasted-call errors"
+        else
+            log_warn "Session was stale-killed by heartbeat monitor (no progress for too long)"
+        fi
     fi
 
     local was_infra_error=false
@@ -177,15 +175,16 @@ spawn_and_wait_session() {
     project_root="$PROJECT_ROOT"
     local default_branch="$base_branch"
 
-    if [[ $exit_code -eq 124 ]]; then
-        log_warn "Session timed out after ${SESSION_TIMEOUT}s"
-        session_status="timed_out"
-    elif [[ "$was_infra_error" == true ]]; then
+    if [[ "$was_infra_error" == true ]]; then
         log_warn "Session failed due to AI CLI/provider infrastructure error"
         log_warn "Infrastructure errors are retried without consuming code retry budget"
         session_status="infra_error"
     elif [[ "$was_stale_killed" == true ]]; then
-        log_warn "Session stale-killed (no progress for ${STALE_KILL_THRESHOLD}s)"
+        if [[ "$stale_kill_reason" == "error_loop" ]]; then
+            log_warn "Session killed due to repeated read-offset/wasted-call error loop"
+        else
+            log_warn "Session stale-killed (no progress for ${STALE_KILL_THRESHOLD}s)"
+        fi
         log_warn "Stale-killed sessions are treated as failed; dev branch is preserved for inspection"
         session_status="crashed"
     elif [[ $exit_code -ne 0 ]]; then
@@ -202,10 +201,10 @@ spawn_and_wait_session() {
             session_status="success"
         else
             local uncommitted=""
-            uncommitted=$(git -C "$project_root" status --porcelain 2>/dev/null | head -1 || true)
+            uncommitted=$(prizm_git_status_safe "$project_root" | head -1 || true)
             if [[ -n "$uncommitted" ]]; then
                 log_warn "Session exited cleanly but produced no commits (uncommitted changes found) — auto-committing..."
-                git -C "$project_root" add -A 2>/dev/null || true
+                prizm_git_add_all_safe "$project_root" || true
                 if git -C "$project_root" commit --no-verify -m "chore($refactor_id): auto-commit session work" 2>/dev/null; then
                     log_info "Auto-commit succeeded"
                     session_status="success"
@@ -224,10 +223,10 @@ spawn_and_wait_session() {
     if [[ "$session_status" == "success" ]]; then
         if git -C "$project_root" rev-parse --is-inside-work-tree >/dev/null 2>&1; then
             local dirty_files=""
-            dirty_files=$(git -C "$project_root" status --porcelain 2>/dev/null || true)
+            dirty_files=$(prizm_git_status_safe "$project_root" || true)
             if [[ -n "$dirty_files" ]]; then
                 log_info "Auto-committing remaining session artifacts..."
-                git -C "$project_root" add -A 2>/dev/null || true
+                prizm_git_add_all_safe "$project_root" || true
                 git -C "$project_root" commit --no-verify --amend --no-edit 2>/dev/null \
                     || git -C "$project_root" commit --no-verify -m "chore($refactor_id): include remaining session artifacts" 2>/dev/null \
                     || true
@@ -305,7 +304,20 @@ sys.exit(0)
         fi
     fi
 
-    # Update refactor status (do NOT commit on dev branch — commit happens after merge)
+    # Final status updates happen after the caller returns to the original branch.
+    _SPAWN_RESULT="$session_status"
+}
+
+finalize_refactor_status_after_branch_return() {
+    local refactor_id="$1"
+    local refactor_list="$2"
+    local session_id="$3"
+    local session_status="$4"
+    local max_retries="$5"
+    local max_infra_retries="$6"
+
+    # Update refactor status on the original branch. The caller commits the
+    # resulting refactor-list diff immediately after this helper returns.
     local update_output
     update_output=$(python3 "$SCRIPTS_DIR/update-refactor-status.py" \
         --refactor-list "$refactor_list" \
@@ -314,14 +326,13 @@ sys.exit(0)
         --session-status "$session_status" \
         --session-id "$session_id" \
         --max-retries "$max_retries" \
+        --max-infra-retries "$max_infra_retries" \
         --action update 2>&1) || {
         log_error "Failed to update refactor status: $update_output"
         update_output=""
     }
 
     _SPAWN_ITEM_STATUS="$(printf '%s' "$update_output" | prizm_extract_update_new_status)"
-
-    _SPAWN_RESULT="$session_status"
 }
 
 # ============================================================
@@ -361,6 +372,8 @@ for item in data.get('refactors', []):
                 --state-dir "$STATE_DIR" \
                 --refactor-id "$_interrupted_id" \
                 --session-status "failed" \
+                --max-retries "$MAX_RETRIES" \
+                --max-infra-retries "$MAX_INFRA_RETRIES" \
                 --action update 2>/dev/null || true
             log_info "Refactor $_interrupted_id marked as failed due to interrupt"
         fi
@@ -391,29 +404,6 @@ check_dependencies() {
     prizm_ensure_git_repo "$_project_root"
 }
 
-run_log_cleanup() {
-    if [[ "$LOG_CLEANUP_ENABLED" != "1" ]]; then
-        return 0
-    fi
-
-    local cleanup_result
-    cleanup_result=$(python3 "$SCRIPTS_DIR/cleanup-logs.py" \
-        --state-dir "$STATE_DIR" \
-        --retention-days "$LOG_RETENTION_DAYS" \
-        --max-total-mb "$LOG_MAX_TOTAL_MB" 2>/dev/null) || {
-        log_warn "Log cleanup failed (continuing)"
-        return 0
-    }
-
-    local deleted reclaimed_kb
-    deleted=$(echo "$cleanup_result" | python3 -c "import sys,json; print(json.load(sys.stdin).get('deleted_files', 0))" 2>/dev/null || echo "0")
-    reclaimed_kb=$(echo "$cleanup_result" | python3 -c "import sys,json; print(int(json.load(sys.stdin).get('reclaimed_bytes', 0)/1024))" 2>/dev/null || echo "0")
-
-    if [[ "$deleted" -gt 0 ]]; then
-        log_info "Log cleanup: deleted $deleted files, reclaimed ${reclaimed_kb}KB"
-    fi
-}
-
 # ============================================================
 # run-one: Run a single refactor
 # ============================================================
@@ -432,7 +422,16 @@ run_one() {
             --dry-run) dry_run=true; shift ;;
             --clean) do_clean=true; shift ;;
             --no-reset) no_reset=true; shift ;;
-            --timeout) shift; SESSION_TIMEOUT="${1:-0}"; shift ;;
+            --max-infra-retries)
+                shift
+                if [[ $# -eq 0 ]]; then
+                    log_error "--max-infra-retries requires a value"
+                    exit 1
+                fi
+                prizm_require_positive_int "--max-infra-retries" "$1"
+                MAX_INFRA_RETRIES="$1"
+                shift
+                ;;
             --mode)
                 shift
                 if [[ $# -eq 0 ]]; then
@@ -473,7 +472,6 @@ run_one() {
     fi
 
     check_dependencies
-    run_log_cleanup
 
     # Initialize state if needed
     if [[ ! -f "$STATE_DIR/pipeline.json" ]]; then
@@ -725,6 +723,8 @@ else:
                 --state-dir "$STATE_DIR" \
                 --refactor-id "$refactor_id" \
                 --session-status "failed" \
+                --max-retries "$MAX_RETRIES" \
+                --max-infra-retries "$MAX_INFRA_RETRIES" \
                 --action update 2>/dev/null || true
             log_info "Refactor $refactor_id marked as failed due to interrupt"
         fi
@@ -758,12 +758,13 @@ else:
     if branch_create "$_proj_root" "$_branch_name" "$_source_branch"; then
         _DEV_BRANCH_NAME="$_branch_name"
     else
-        log_warn "Failed to create branch; running session on current branch"
+        log_error "Failed to create branch: $_branch_name from $_source_branch — aborting to avoid running on $_source_branch"
+        return 1
     fi
 
     spawn_and_wait_session \
         "$refactor_id" "$refactor_list" "$session_id" \
-        "$bootstrap_prompt" "$session_dir" 999 "$refactor_model" "$_ORIGINAL_BRANCH"
+        "$bootstrap_prompt" "$session_dir" "$MAX_RETRIES" "$MAX_INFRA_RETRIES" "$refactor_model" "$_ORIGINAL_BRANCH"
     local session_status="$_SPAWN_RESULT"
 
     # Merge dev branch back to original on success
@@ -773,7 +774,7 @@ else:
         else
             log_warn "Auto-merge failed — dev branch preserved: $_DEV_BRANCH_NAME"
             log_warn "Merge manually: git checkout $_ORIGINAL_BRANCH && git rebase $_DEV_BRANCH_NAME"
-            _DEV_BRANCH_NAME=""
+            session_status="merge_conflict"
         fi
     elif [[ -n "$_DEV_BRANCH_NAME" ]]; then
         # Session failed — preserve dev branch for inspection
@@ -784,10 +785,12 @@ else:
     # GUARANTEED: always return to original branch regardless of success/failure/merge outcome
     branch_ensure_return "$_proj_root" "$_ORIGINAL_BRANCH"
 
+    finalize_refactor_status_after_branch_return \
+        "$refactor_id" "$refactor_list" "$session_id" "$session_status" "$MAX_RETRIES" "$MAX_INFRA_RETRIES"
+
     # Commit refactor status update on the original branch (after guaranteed return)
     if ! git -C "$_proj_root" diff --quiet "$refactor_list" 2>/dev/null; then
-        git -C "$_proj_root" add "$refactor_list"
-        git -C "$_proj_root" commit --no-verify -m "chore($refactor_id): update refactor status" 2>/dev/null || true
+        prizm_git_commit_paths "$_proj_root" "chore($refactor_id): update refactor status" "$refactor_list"
     fi
 
     echo ""
@@ -808,7 +811,41 @@ else:
 # ============================================================
 
 main() {
-    local refactor_list="${1:-$PRIZMKIT_DIR/plans/refactor-list.json}"
+    local refactor_list="$PRIZMKIT_DIR/plans/refactor-list.json"
+
+    while [[ $# -gt 0 ]]; do
+        case "$1" in
+            --max-infra-retries)
+                shift
+                if [[ $# -eq 0 ]]; then
+                    log_error "--max-infra-retries requires a value"
+                    exit 1
+                fi
+                prizm_require_positive_int "--max-infra-retries" "$1"
+                MAX_INFRA_RETRIES="$1"
+                shift
+                ;;
+            --)
+                shift
+                break
+                ;;
+            -*)
+                log_error "Unknown option: $1"
+                show_help
+                exit 1
+                ;;
+            *)
+                refactor_list="$1"
+                shift
+                ;;
+        esac
+    done
+
+    if [[ $# -gt 0 ]]; then
+        log_error "Unexpected arguments: $*"
+        show_help
+        exit 1
+    fi
 
     if [[ ! "$refactor_list" = /* ]]; then
         refactor_list="$(pwd)/$refactor_list"
@@ -823,7 +860,6 @@ main() {
     fi
 
     check_dependencies
-    run_log_cleanup
 
     # Initialize pipeline state if needed
     if [[ ! -f "$STATE_DIR/pipeline.json" ]]; then
@@ -870,12 +906,8 @@ main() {
     echo -e "${BOLD}        Refactor Pipeline Runner Started${NC}"
     echo -e "${BOLD}════════════════════════════════════════════════════${NC}"
     log_info "Refactor list: $refactor_list"
-    log_info "Max retries per refactor: $MAX_RETRIES"
-    if [[ $SESSION_TIMEOUT -gt 0 ]]; then
-        log_info "Session timeout: ${SESSION_TIMEOUT}s"
-    else
-        log_info "Session timeout: none"
-    fi
+    log_info "Max code retries per refactor: $MAX_RETRIES"
+    log_info "Max infrastructure retries per refactor: $MAX_INFRA_RETRIES"
     log_info "AI CLI: $CLI_CMD (platform: $PLATFORM)"
     if [[ -n "${MODEL:-}" ]]; then
         log_info "Default Model: $MODEL"
@@ -920,6 +952,7 @@ main() {
             --refactor-list "$refactor_list" \
             --state-dir "$STATE_DIR" \
             --max-retries "$MAX_RETRIES" \
+            --max-infra-retries "$MAX_INFRA_RETRIES" \
             --action get_next 2>/dev/null); then
             log_error "Failed to get next refactor"
             break
@@ -1007,27 +1040,40 @@ DEPLOY_PROMPT_EOF
         fi
 
         # Parse refactor info
-        local refactor_id refactor_title refactor_complexity retry_count resume_phase
+        local refactor_id refactor_title refactor_complexity retry_count infra_error_count resume_phase
         refactor_id=$(echo "$next_refactor" | jq -r '.refactor_id')
         refactor_title=$(echo "$next_refactor" | jq -r '.title')
         refactor_complexity=$(echo "$next_refactor" | jq -r '.complexity')
         retry_count=$(echo "$next_refactor" | jq -r '.retry_count // 0')
+        infra_error_count=$(echo "$next_refactor" | jq -r '.infra_error_count // 0')
         resume_phase=$(echo "$next_refactor" | jq -r '.resume_from_phase // "null"')
 
         echo ""
         echo -e "${BOLD}────────────────────────────────────────────────────${NC}"
         log_info "Refactor: ${BOLD}$refactor_id${NC} — $refactor_title"
-        log_info "Complexity: $refactor_complexity | Retry: $retry_count / $MAX_RETRIES"
+        log_info "Complexity: $refactor_complexity"
+        log_info "Code retry: $retry_count / $MAX_RETRIES"
+        log_info "Infrastructure retry: $infra_error_count / $MAX_INFRA_RETRIES"
         if [[ "$resume_phase" != "null" ]]; then
             log_info "Resuming from Phase $resume_phase"
         fi
         echo -e "${BOLD}────────────────────────────────────────────────────${NC}"
 
-        # Pre-commit any dirty tree from previous iteration
-        if ! git -C "$_proj_root" diff --quiet HEAD 2>/dev/null || [ -n "$(git -C "$_proj_root" ls-files --others --exclude-standard 2>/dev/null)" ]; then
-            log_info "Dirty working tree detected — committing before $refactor_id..."
-            git -C "$_proj_root" add -A 2>/dev/null || true
-            git -C "$_proj_root" commit --no-verify -m "chore: capture artifacts before $refactor_id session" 2>/dev/null || true
+        # Auto-commit any uncommitted workspace changes before starting the pipeline.
+        # This ensures a clean baseline without blocking execution.
+        local _dirty_files=""
+        _dirty_files=$(prizm_git_status_non_bookkeeping "$_proj_root" "$STATE_DIR" "$refactor_list" || true)
+        if [[ -n "$_dirty_files" ]]; then
+            log_info "Auto-committing workspace changes before $refactor_id..."
+            prizm_git_add_all_safe "$_proj_root"
+            git -C "$_proj_root" commit --no-verify -m "chore: workspace snapshot ready for run $refactor_id" 2>/dev/null || true
+        fi
+
+        local _bookkeeping_dirty=""
+        _bookkeeping_dirty=$(prizm_git_status_bookkeeping "$_proj_root" "$STATE_DIR" "$refactor_list" || true)
+        if [[ -n "$_bookkeeping_dirty" ]]; then
+            log_info "Committing pending pipeline bookkeeping before $refactor_id..."
+            prizm_git_commit_paths "$_proj_root" "chore($refactor_id): include pipeline bookkeeping" "$refactor_list" "$STATE_DIR"
         fi
 
         # Mark refactor as in-progress BEFORE creating dev branch
@@ -1040,8 +1086,7 @@ DEPLOY_PROMPT_EOF
             --action start >/dev/null 2>&1 || true
         # Commit the in_progress status on the original branch
         if ! git -C "$_proj_root" diff --quiet "$refactor_list" 2>/dev/null; then
-            git -C "$_proj_root" add "$refactor_list" 2>/dev/null || true
-            git -C "$_proj_root" commit --no-verify -m "chore($refactor_id): mark in_progress" 2>/dev/null || true
+            prizm_git_commit_paths "$_proj_root" "chore($refactor_id): mark in_progress" "$refactor_list"
         fi
 
         # Create per-refactor dev branch (from the now-updated original branch)
@@ -1050,8 +1095,9 @@ DEPLOY_PROMPT_EOF
             _DEV_BRANCH_NAME="$_refactor_branch"
             log_info "Dev branch: $_refactor_branch"
         else
-            log_warn "Failed to create dev branch; running on current branch: $_ORIGINAL_BRANCH"
+            log_error "Failed to create dev branch: $_refactor_branch from $_ORIGINAL_BRANCH — aborting to avoid running on $_ORIGINAL_BRANCH"
             _DEV_BRANCH_NAME=""
+            break
         fi
 
         # Generate session
@@ -1122,7 +1168,7 @@ DEPLOY_PROMPT_EOF
 
         spawn_and_wait_session \
             "$refactor_id" "$refactor_list" "$session_id" \
-            "$bootstrap_prompt" "$session_dir" "$MAX_RETRIES" "$refactor_model" "$_ORIGINAL_BRANCH"
+            "$bootstrap_prompt" "$session_dir" "$MAX_RETRIES" "$MAX_INFRA_RETRIES" "$refactor_model" "$_ORIGINAL_BRANCH"
 
         # Validate key artifacts after successful session
         if [[ "$_SPAWN_RESULT" == "success" ]]; then
@@ -1135,7 +1181,6 @@ DEPLOY_PROMPT_EOF
         fi
 
         local session_status="$_SPAWN_RESULT"
-        local item_status_after_session="${_SPAWN_ITEM_STATUS:-}"
 
         # Merge per-refactor dev branch back to original on success
         if [[ "$session_status" == "success" && -n "$_DEV_BRANCH_NAME" ]]; then
@@ -1144,7 +1189,7 @@ DEPLOY_PROMPT_EOF
             else
                 log_warn "Auto-merge failed — dev branch preserved: $_DEV_BRANCH_NAME"
                 log_warn "Merge manually: git checkout $_ORIGINAL_BRANCH && git rebase $_DEV_BRANCH_NAME"
-                _DEV_BRANCH_NAME=""
+                session_status="merge_conflict"
             fi
         elif [[ -n "$_DEV_BRANCH_NAME" ]]; then
             # Session failed — preserve dev branch for inspection
@@ -1155,10 +1200,13 @@ DEPLOY_PROMPT_EOF
         # GUARANTEED: always return to original branch regardless of success/failure/merge outcome
         branch_ensure_return "$_proj_root" "$_ORIGINAL_BRANCH"
 
+        finalize_refactor_status_after_branch_return \
+            "$refactor_id" "$refactor_list" "$session_id" "$session_status" "$MAX_RETRIES" "$MAX_INFRA_RETRIES"
+        local item_status_after_session="${_SPAWN_ITEM_STATUS:-}"
+
         # Commit refactor status update on the original branch (after guaranteed return)
         if ! git -C "$_proj_root" diff --quiet "$refactor_list" 2>/dev/null; then
-            git -C "$_proj_root" add "$refactor_list"
-            git -C "$_proj_root" commit --no-verify -m "chore($refactor_id): update refactor status" 2>/dev/null || true
+            prizm_git_commit_paths "$_proj_root" "chore($refactor_id): update refactor status" "$refactor_list"
         fi
 
         # Stuck detection
@@ -1185,7 +1233,7 @@ DEPLOY_PROMPT_EOF
             log_error "════════════════════════════════════════════════════"
             break
         elif [[ "$session_status" != "success" && "$STOP_ON_FAILURE" == "1" ]]; then
-            log_info "STOP_ON_FAILURE: $refactor_id is ${item_status_after_session:-unknown}; retry budget not exhausted, continuing."
+            log_info "STOP_ON_FAILURE: $refactor_id is ${item_status_after_session:-unknown}; code/infra retry budget not exhausted, continuing."
         fi
 
         log_info "Pausing 5s before next refactor..."
@@ -1201,7 +1249,7 @@ show_help() {
     echo "Usage: $0 <command> [options]"
     echo ""
     echo "Commands:"
-    echo "  run [.prizmkit/plans/refactor-list.json]                Run all refactors in dependency-topological order"
+    echo "  run [.prizmkit/plans/refactor-list.json] [options]       Run all refactors in dependency-topological order"
     echo "  run <refactor-id> [options]              Run a single refactor"
     echo "  status [.prizmkit/plans/refactor-list.json]              Show refactor pipeline status"
     echo "  reset                                    Clear all refactor state"
@@ -1211,14 +1259,17 @@ show_help() {
     echo "  --dry-run                   Generate bootstrap prompt only, don't spawn session"
     echo "  --clean                     Delete artifacts and reset before running"
     echo "  --no-reset                  Skip status reset (preserve retry count)"
-    echo "  --timeout N                 Session timeout in seconds (default: 0 = no limit)"
+    echo "  --max-infra-retries N       Max infrastructure/provider retries (default: 3)"
     echo "  --mode <lite|standard|full> Override pipeline mode"
     echo "  --critic                    Enable adversarial critic review"
     echo "  --no-critic                 Disable adversarial critic review"
     echo ""
+    echo "Pipeline Options (run all):"
+    echo "  --max-infra-retries N       Max infrastructure/provider retries (default: 3)"
+    echo ""
     echo "Environment Variables:"
-    echo "  MAX_RETRIES              Max retries per refactor (default: 3)"
-    echo "  SESSION_TIMEOUT          Session timeout in seconds (default: 0 = no limit)"
+    echo "  MAX_RETRIES              Max code retries per refactor (default: 3)"
+    echo "  MAX_INFRA_RETRIES     Max infrastructure/provider retries per refactor (default: 3)"
     echo "  MODEL                    Default AI model (overridden by per-refactor model in refactor list)"
     echo "  PRIZMKIT_EFFORT          AI reasoning effort (low|medium|high|xhigh|max; max is Claude Code only)"
     echo "  PIPELINE_MODE            Default pipeline mode: lite|standard|full (overridden by --mode)"
@@ -1229,15 +1280,11 @@ show_help() {
     echo "  HEARTBEAT_INTERVAL       Heartbeat log interval in seconds (default: 30)"
     echo "  STALE_KILL_THRESHOLD     Auto-kill session after N seconds of no progress (default: 900)"
     echo "  STOP_ON_FAILURE          Stop pipeline when a task exhausts retries: 0|1 (default: 0)"
-    echo "  LOG_CLEANUP_ENABLED      Run log cleanup before execution (default: 1)"
-    echo "  LOG_RETENTION_DAYS       Delete logs older than N days (default: 14)"
-    echo "  LOG_MAX_TOTAL_MB         Keep total logs under N MB (default: 1024)"
     echo ""
     echo "Examples:"
     echo "  ./run-refactor.sh run                                    # Run all refactors"
     echo "  ./run-refactor.sh run .prizmkit/plans/refactor-list.json                 # Custom refactor list"
     echo "  ./run-refactor.sh run R-001 --dry-run                    # Inspect generated prompt"
-    echo "  ./run-refactor.sh run R-001 --timeout 3600               # 1h timeout"
     echo "  ./run-refactor.sh status                                 # Show status"
     echo "  STRICT_BEHAVIOR_CHECK=0 ./run-refactor.sh run            # Skip full test suite"
     echo "  MAX_RETRIES=5 ./run-refactor.sh run                      # Custom retries"
@@ -1249,7 +1296,7 @@ case "${1:-run}" in
         if [[ "${1:-}" =~ ^[Rr]-[0-9]+ ]]; then
             run_one "$@"
         else
-            main "${1:-$PRIZMKIT_DIR/plans/refactor-list.json}"
+            main "$@"
         fi
         ;;
     status)
@@ -1261,6 +1308,8 @@ case "${1:-run}" in
         python3 "$SCRIPTS_DIR/update-refactor-status.py" \
             --refactor-list "${2:-$PRIZMKIT_DIR/plans/refactor-list.json}" \
             --state-dir "$STATE_DIR" \
+            --max-retries "$MAX_RETRIES" \
+            --max-infra-retries "$MAX_INFRA_RETRIES" \
             --action status
         ;;
     reset)