@indigoai-us/hq-cli 5.77.14 → 5.78.0

package/assets/scaffold/core/scripts/detect-stale-core-policy-mirror.sh

@@ -0,0 +1,185 @@
+#!/usr/bin/env bash
+# detect-stale-core-policy-mirror.sh — find leftover core/policies/ copies of a
+# personal/policies/ twin from before the personal→core policy mirror was
+# retired.
+#
+# Background: HQ used to symlink personal/policies/<slug>.md into
+# core/policies/<slug>.md at reindex time. That mirror is retired — the policy
+# trigger hook (.claude/hooks/inject-policy-on-trigger.sh) now reads
+# personal/policies/ DIRECTLY, ahead of core/policies/, and dedups by policy id.
+# But an install upgraded across the retirement can still carry stale REGULAR
+# FILE copies of personal policies inside core/policies/. Those copies are
+# release-owned scaffold: /update-hq will not refresh them from personal/, and a
+# copy that is missing the when:/on: trigger frontmatter its personal twin has
+# gained (via migrate-policy-triggers.sh) can shadow or diverge from the live
+# rule in confusing ways.
+#
+# This tool classifies every core/policies twin of a personal policy:
+#   identical  — byte-identical to the personal twin: an orphaned mirror copy,
+#                behavior-neutral to remove (the personal twin still loads, and
+#                /update-hq restores any genuinely release-shipped file).
+#   symlink    — a leftover personal→core mirror symlink: resolves back to
+#                personal, so also a behavior-neutral orphan.
+#   diverged   — content differs from the personal twin. DRIFT CAN RUN EITHER
+#                DIRECTION (the core copy may be the newer, genericized one).
+#                These are NEVER pruned — a human must classify each.
+#
+# Modes:
+#   (default)          human-readable report; always exit 0.
+#   --json             machine-readable JSON array; exit 0.
+#   --check            exit 0 if no twins, 1 if any twin exists (advisory gate).
+#   --prune-identical  remove ONLY identical + leftover-symlink orphans; never
+#                      touch diverged; print each action; exit 0.
+#
+# It never writes to personal/, never syncs personal→core, and never removes a
+# core/policies file that has no personal twin (release-shipped policies are
+# left untouched by construction).
+
+set -euo pipefail
+
+MODE="report"
+case "${1:-}" in
+  "")                 MODE="report" ;;
+  --json)             MODE="json" ;;
+  --check)            MODE="check" ;;
+  --prune-identical)  MODE="prune" ;;
+  -h|--help)
+    sed -n '2,45p' "$0"
+    exit 0 ;;
+  *)
+    printf 'detect-stale-core-policy-mirror: unknown argument: %s\n' "$1" >&2
+    printf 'usage: %s [--json|--check|--prune-identical]\n' "$0" >&2
+    exit 2 ;;
+esac
+
+# Resolve HQ root: explicit env wins, else the tree this script ships in.
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+HQ_ROOT="${HQ_ROOT:-${CLAUDE_PROJECT_DIR:-$(cd "$SCRIPT_DIR/../.." && pwd)}}"
+
+PERSONAL_DIR="$HQ_ROOT/personal/policies"
+CORE_DIR="$HQ_ROOT/core/policies"
+
+# Collect twins into parallel arrays (bash 3.2-safe; no associative arrays).
+names=()
+statuses=()
+
+if [ -d "$PERSONAL_DIR" ]; then
+  for pf in "$PERSONAL_DIR"/*.md; do
+    [ -e "$pf" ] || continue          # no-glob-match guard (nullglob-free)
+    base="$(basename "$pf")"
+    case "$base" in example-policy.md|README.md) continue ;; esac
+    cf="$CORE_DIR/$base"
+    [ -e "$cf" ] || [ -L "$cf" ] || continue   # no core twin → nothing to flag
+    if [ -L "$cf" ]; then
+      status="symlink"
+    elif cmp -s "$pf" "$cf"; then
+      status="identical"
+    else
+      status="diverged"
+    fi
+    names+=("$base")
+    statuses+=("$status")
+  done
+fi
+
+total="${#names[@]}"
+
+# Tally.
+n_identical=0; n_symlink=0; n_diverged=0
+i=0
+while [ "$i" -lt "$total" ]; do
+  case "${statuses[$i]}" in
+    identical) n_identical=$((n_identical + 1)) ;;
+    symlink)   n_symlink=$((n_symlink + 1)) ;;
+    diverged)  n_diverged=$((n_diverged + 1)) ;;
+  esac
+  i=$((i + 1))
+done
+n_orphan=$((n_identical + n_symlink))
+
+json_escape() { printf '%s' "$1" | sed 's/\\/\\\\/g; s/"/\\"/g'; }
+
+emit_json() {
+  printf '['
+  local i=0 sep=""
+  while [ "$i" -lt "$total" ]; do
+    printf '%s{"policy":"%s","status":"%s","core_path":"core/policies/%s","personal_path":"personal/policies/%s"}' \
+      "$sep" "$(json_escape "${names[$i]}")" "${statuses[$i]}" \
+      "$(json_escape "${names[$i]}")" "$(json_escape "${names[$i]}")"
+    sep=","
+    i=$((i + 1))
+  done
+  printf ']\n'
+}
+
+case "$MODE" in
+  json)
+    emit_json
+    exit 0 ;;
+
+  check)
+    if [ "$total" -eq 0 ]; then
+      echo "detect-stale-core-policy-mirror: clean (no core/policies twins of a personal policy)"
+      exit 0
+    fi
+    printf 'detect-stale-core-policy-mirror: %s twin(s) found — %s prunable orphan(s), %s diverged (need human classification)\n' \
+      "$total" "$n_orphan" "$n_diverged" >&2
+    exit 1 ;;
+
+  prune)
+    if [ "$total" -eq 0 ]; then
+      echo "detect-stale-core-policy-mirror: nothing to prune (no twins)"
+      exit 0
+    fi
+    removed=0; skipped=0
+    i=0
+    while [ "$i" -lt "$total" ]; do
+      base="${names[$i]}"; st="${statuses[$i]}"
+      case "$st" in
+        identical|symlink)
+          rm -f "$CORE_DIR/$base"
+          printf 'pruned  (%s): core/policies/%s\n' "$st" "$base"
+          removed=$((removed + 1)) ;;
+        diverged)
+          printf 'SKIPPED (diverged, needs human classification): core/policies/%s\n' "$base"
+          skipped=$((skipped + 1)) ;;
+      esac
+      i=$((i + 1))
+    done
+    printf 'detect-stale-core-policy-mirror: pruned %s orphan(s); left %s diverged twin(s) for review.\n' \
+      "$removed" "$skipped"
+    exit 0 ;;
+
+  report|*)
+    if [ "$total" -eq 0 ]; then
+      echo "detect-stale-core-policy-mirror: clean — no core/policies copies shadow a personal twin."
+      exit 0
+    fi
+    printf 'Stale personal→core policy mirror copies in core/policies/ (%s twin(s)):\n\n' "$total"
+    if [ "$n_orphan" -gt 0 ]; then
+      printf 'Orphaned mirror copies (behavior-neutral — remove with --prune-identical):\n'
+      i=0
+      while [ "$i" -lt "$total" ]; do
+        case "${statuses[$i]}" in
+          identical) printf '  identical  core/policies/%s\n' "${names[$i]}" ;;
+          symlink)   printf '  symlink    core/policies/%s\n' "${names[$i]}" ;;
+        esac
+        i=$((i + 1))
+      done
+      printf '\n'
+    fi
+    if [ "$n_diverged" -gt 0 ]; then
+      printf 'Diverged copies (DRIFT MAY RUN EITHER DIRECTION — classify by hand, do NOT blind-sync):\n'
+      i=0
+      while [ "$i" -lt "$total" ]; do
+        if [ "${statuses[$i]}" = "diverged" ]; then
+          printf '  diverged   core/policies/%s   (diff: git diff --no-index personal/policies/%s core/policies/%s)\n' \
+            "${names[$i]}" "${names[$i]}" "${names[$i]}"
+        fi
+        i=$((i + 1))
+      done
+      printf '\n'
+    fi
+    printf 'Summary: %s orphan(s), %s diverged.\n' "$n_orphan" "$n_diverged"
+    exit 0 ;;
+esac

package/assets/scaffold/core/scripts/hq-status-summary.sh

@@ -0,0 +1,187 @@
+#!/usr/bin/env bash
+# hq-status-summary.sh — classify noisy HQ git status output.
+#
+# This is deliberately read-only. It turns `git status --porcelain --ignored`
+# into a compact summary that separates current-session touched paths from the
+# local HQ baseline noise that may remain untracked on installer-created roots.
+
+set -euo pipefail
+
+ROOT="${HQ_ROOT:-$(git rev-parse --show-toplevel 2>/dev/null || pwd)}"
+PORCELAIN_FILE=""
+SESSION_FILES_JSON="[]"
+SESSION_FILES_FILE=""
+FORMAT="json"
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --porcelain-file) PORCELAIN_FILE="$2"; shift 2 ;;
+    --session-files-json) SESSION_FILES_JSON="$2"; shift 2 ;;
+    --session-files-file) SESSION_FILES_FILE="$2"; shift 2 ;;
+    --json) FORMAT="json"; shift ;;
+    --text) FORMAT="text"; shift ;;
+    *) echo "Unknown arg: $1" >&2; exit 2 ;;
+  esac
+done
+
+# --session-files-file <path> keeps a large session changeset off argv (mirrors
+# the existing --porcelain-file pattern). A missing path is a hard error rather
+# than a silent empty summary. It is resolved before the cd below because a
+# relative path is relative to the caller's cwd, not ROOT.
+if [[ -n "$SESSION_FILES_FILE" ]]; then
+  if [[ ! -f "$SESSION_FILES_FILE" ]]; then
+    echo "hq-status-summary: --session-files-file not found: $SESSION_FILES_FILE" >&2
+    exit 3
+  fi
+  case "$SESSION_FILES_FILE" in
+    /*) : ;;
+    *) SESSION_FILES_FILE="$PWD/$SESSION_FILES_FILE" ;;
+  esac
+fi
+
+cd "$ROOT"
+
+tmpdir=$(mktemp -d)
+trap 'rm -rf "$tmpdir"' EXIT
+
+tracked="$tmpdir/tracked"
+session_untracked="$tmpdir/session_untracked"
+baseline_untracked="$tmpdir/baseline_untracked"
+unrelated_untracked="$tmpdir/unrelated_untracked"
+ignored="$tmpdir/ignored"
+unsafe_session="$tmpdir/unsafe_session"
+: >"$tracked"; : >"$session_untracked"; : >"$baseline_untracked"; : >"$unrelated_untracked"; : >"$ignored"; : >"$unsafe_session"
+
+session_paths_file="$tmpdir/session_paths"
+session_files_jq='
+  if type != "array" then empty else
+    .[] | if type == "string" then . else (.path // empty) end
+  end
+'
+if [[ -n "$SESSION_FILES_FILE" ]]; then
+  jq -r "$session_files_jq" "$SESSION_FILES_FILE" 2>/dev/null | sed '/^$/d' > "$session_paths_file" || : > "$session_paths_file"
+else
+  printf '%s' "$SESSION_FILES_JSON" | jq -r "$session_files_jq" 2>/dev/null | sed '/^$/d' > "$session_paths_file" || : > "$session_paths_file"
+fi
+
+baseline_patterns_file="$tmpdir/baseline_patterns"
+if [[ -f workspace/baseline/hq-local-baseline.json ]]; then
+  jq -r '.categories[]?.patterns[]?' workspace/baseline/hq-local-baseline.json 2>/dev/null > "$baseline_patterns_file" || : > "$baseline_patterns_file"
+else
+  cat > "$baseline_patterns_file" <<'PATTERNS'
+companies/*
+workspace/*
+repos/*
+core/settings/*
+.hq/*
+PATTERNS
+fi
+
+is_session_path() {
+  local path="$1"
+  local p
+  while IFS= read -r p; do
+    [[ -z "$p" ]] && continue
+    if [[ "$path" == "$p" || "$path" == "$p/"* || "$p" == "$path/"* ]]; then
+      return 0
+    fi
+  done < "$session_paths_file"
+  return 1
+}
+
+is_baseline_path() {
+  local path="$1"
+  local pat
+  while IFS= read -r pat; do
+    [[ -z "$pat" ]] && continue
+    case "$path" in
+      $pat) return 0 ;;
+    esac
+  done < "$baseline_patterns_file"
+  return 1
+}
+
+status_source="$tmpdir/porcelain"
+if [[ -n "$PORCELAIN_FILE" ]]; then
+  cp "$PORCELAIN_FILE" "$status_source"
+else
+  git status --porcelain --ignored > "$status_source"
+fi
+
+while IFS= read -r line; do
+  [[ -z "$line" ]] && continue
+  code="${line:0:2}"
+  path="${line:3}"
+  if [[ "$path" == *" -> "* ]]; then
+    path="${path##* -> }"
+  fi
+
+  case "$code" in
+    '!!')
+      printf '%s\n' "$path" >> "$ignored"
+      ;;
+    '??')
+      if is_session_path "$path"; then
+        printf '%s\n' "$path" >> "$session_untracked"
+      elif is_baseline_path "$path"; then
+        printf '%s\n' "$path" >> "$baseline_untracked"
+      else
+        printf '%s\n' "$path" >> "$unrelated_untracked"
+      fi
+      ;;
+    *)
+      printf '%s\n' "$path" >> "$tracked"
+      ;;
+  esac
+done < "$status_source"
+
+# Materialize each classified list as a JSON array file, then feed jq via
+# --slurpfile. --argjson puts the whole array on argv and overflows CreateProcess
+# (~32KB) / MAX_ARG_STRLEN (128KB) when thousands of untracked paths classify here.
+write_json_array() {
+  jq -R -s 'split("\n") | map(select(length > 0))' "$1" > "$2"
+}
+
+tracked_json="$tmpdir/tracked.json"
+session_untracked_json="$tmpdir/session_untracked.json"
+baseline_untracked_json="$tmpdir/baseline_untracked.json"
+unrelated_untracked_json="$tmpdir/unrelated_untracked.json"
+ignored_json="$tmpdir/ignored.json"
+
+write_json_array "$tracked" "$tracked_json"
+write_json_array "$session_untracked" "$session_untracked_json"
+write_json_array "$baseline_untracked" "$baseline_untracked_json"
+write_json_array "$unrelated_untracked" "$unrelated_untracked_json"
+write_json_array "$ignored" "$ignored_json"
+
+if [[ "$FORMAT" == "text" ]]; then
+  printf 'tracked_changes=%s\n' "$(jq 'length' "$tracked_json")"
+  printf 'session_untracked=%s\n' "$(jq 'length' "$session_untracked_json")"
+  printf 'baseline_untracked=%s\n' "$(jq 'length' "$baseline_untracked_json")"
+  printf 'unrelated_untracked=%s\n' "$(jq 'length' "$unrelated_untracked_json")"
+  printf 'ignored=%s\n' "$(jq 'length' "$ignored_json")"
+else
+  jq -n \
+    --slurpfile tracked "$tracked_json" \
+    --slurpfile session_untracked "$session_untracked_json" \
+    --slurpfile baseline_untracked "$baseline_untracked_json" \
+    --slurpfile unrelated_untracked "$unrelated_untracked_json" \
+    --slurpfile ignored "$ignored_json" \
+    '{
+      counts: {
+        tracked_changes: ($tracked[0] | length),
+        session_touched_untracked: ($session_untracked[0] | length),
+        baseline_untracked: ($baseline_untracked[0] | length),
+        unrelated_untracked: ($unrelated_untracked[0] | length),
+        ignored: ($ignored[0] | length),
+        baseline_noise: (($baseline_untracked[0] | length) + ($unrelated_untracked[0] | length))
+      },
+      files: {
+        tracked_changes: $tracked[0],
+        session_touched_untracked: $session_untracked[0],
+        baseline_untracked: $baseline_untracked[0],
+        unrelated_untracked: $unrelated_untracked[0],
+        ignored: $ignored[0]
+      }
+    }'
+fi

package/assets/scaffold/core/scripts/ontology-readme-drift.sh

@@ -0,0 +1,264 @@
+#!/usr/bin/env bash
+# hq-core: public
+# Detect README team-table names that no longer match ontology person entities.
+set -euo pipefail
+
+usage() {
+  echo "USAGE: ontology-readme-drift.sh <company-slug> [--root <HQ_ROOT>]" >&2
+}
+
+SLUG=""
+ROOT_ARG=""
+
+while [ "$#" -gt 0 ]; do
+  case "$1" in
+    --root)
+      if [ "$#" -lt 2 ]; then
+        usage
+        exit 2
+      fi
+      ROOT_ARG="$2"
+      shift 2
+      ;;
+    -*)
+      usage
+      exit 2
+      ;;
+    *)
+      if [ -n "$SLUG" ]; then
+        usage
+        exit 2
+      fi
+      SLUG="$1"
+      shift
+      ;;
+  esac
+done
+
+if [ -z "$SLUG" ]; then
+  usage
+  exit 2
+fi
+
+if [ -n "$ROOT_ARG" ]; then
+  ROOT="$ROOT_ARG"
+elif [ -n "${HQ_ROOT:-}" ]; then
+  ROOT="$HQ_ROOT"
+elif [ -n "${CLAUDE_PROJECT_DIR:-}" ]; then
+  ROOT="$CLAUDE_PROJECT_DIR"
+else
+  SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+  ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
+fi
+
+README="$ROOT/companies/$SLUG/README.md"
+PERSONS_DIR="$ROOT/companies/$SLUG/ontology/entities/person"
+
+if [ ! -f "$README" ]; then
+  echo "ERROR: README missing: companies/$SLUG/README.md" >&2
+  exit 2
+fi
+
+if [ ! -d "$PERSONS_DIR" ]; then
+  echo "ERROR: ontology person directory missing: companies/$SLUG/ontology/entities/person" >&2
+  exit 2
+fi
+
+has_person_file=0
+for person_file in "$PERSONS_DIR"/*.md; do
+  [ -f "$person_file" ] || continue
+  has_person_file=1
+  break
+done
+
+if [ "$has_person_file" -eq 0 ]; then
+  echo "ERROR: ontology person directory is empty: companies/$SLUG/ontology/entities/person" >&2
+  exit 2
+fi
+
+trim_name_awk='
+function trim(s) {
+  sub(/^[[:space:]]+/, "", s)
+  sub(/[[:space:]]+$/, "", s)
+  return s
+}
+'
+
+extract_team_names() {
+  # Parse only the first markdown table after ## Team; column 2 is Name.
+  awk "$trim_name_awk"'
+    /^##[[:space:]]+Team[[:space:]]*$/ {
+      in_team = 1
+      next
+    }
+    in_team && /^##[[:space:]]+/ {
+      exit
+    }
+    in_team {
+      if ($0 ~ /^[[:space:]]*\|/) {
+        in_table = 1
+        row += 1
+        if (row <= 2) {
+          next
+        }
+        line = $0
+        sub(/^[[:space:]]*/, "", line)
+        split(line, cols, /\|/)
+        name = cols[3]
+        gsub(/<!--[^>]*-->/, "", name)
+        name = trim(name)
+        if (name != "") {
+          print name
+        }
+        next
+      }
+      if (in_table) {
+        exit
+      }
+    }
+  ' "$README"
+}
+
+extract_person_names() {
+  # Read canonical_name and aliases from YAML frontmatter only.
+  awk "$trim_name_awk"'
+    function unquote(s) {
+      s = trim(s)
+      if ((s ~ /^".*"$/) || (s ~ /^'\''.*'\''$/)) {
+        s = substr(s, 2, length(s) - 2)
+      }
+      return s
+    }
+    /^---[[:space:]]*$/ {
+      markers += 1
+      if (markers == 1) {
+        in_frontmatter = 1
+        next
+      }
+      if (markers == 2) {
+        exit
+      }
+    }
+    !in_frontmatter {
+      next
+    }
+    /^[[:space:]]*aliases:[[:space:]]*$/ {
+      in_aliases = 1
+      next
+    }
+    in_aliases && /^[[:space:]]*-[[:space:]]*/ {
+      line = $0
+      sub(/^[[:space:]]*-[[:space:]]*/, "", line)
+      line = unquote(line)
+      if (line != "") {
+        print line
+      }
+      next
+    }
+    in_aliases && $0 !~ /^[[:space:]]*($|#)/ {
+      in_aliases = 0
+    }
+    /^[[:space:]]*canonical_name[[:space:]]*:/ {
+      line = $0
+      sub(/^[[:space:]]*canonical_name[[:space:]]*:[[:space:]]*/, "", line)
+      line = unquote(line)
+      if (line != "") {
+        print line
+      }
+      next
+    }
+  ' "$1"
+}
+
+normalize_name() {
+  LC_ALL=C awk -v s="$1" 'BEGIN {
+    s = tolower(s)
+    gsub(/[^[:alnum:][:space:]]/, "", s)
+    gsub(/[[:space:]]+/, " ", s)
+    sub(/^ /, "", s)
+    sub(/ $/, "", s)
+    print s
+  }'
+}
+
+contains_token() {
+  token="$1"
+  shift
+  for candidate in "$@"; do
+    if [ "$candidate" = "$token" ]; then
+      return 0
+    fi
+  done
+  return 1
+}
+
+name_matches() {
+  readme_norm="$1"
+  ontology_norm="$2"
+
+  if [ "$readme_norm" = "$ontology_norm" ]; then
+    return 0
+  fi
+  if [ -z "$readme_norm" ] || [ -z "$ontology_norm" ]; then
+    return 1
+  fi
+
+  IFS=' ' read -r -a readme_tokens <<< "$readme_norm"
+  IFS=' ' read -r -a ontology_tokens <<< "$ontology_norm"
+
+  if [ "${#readme_tokens[@]}" -lt "${#ontology_tokens[@]}" ] ||
+    { [ "${#readme_tokens[@]}" -eq "${#ontology_tokens[@]}" ] && [ "${#readme_norm}" -le "${#ontology_norm}" ]; }; then
+    for token in "${readme_tokens[@]}"; do
+      contains_token "$token" "${ontology_tokens[@]}" || return 1
+    done
+  else
+    for token in "${ontology_tokens[@]}"; do
+      contains_token "$token" "${readme_tokens[@]}" || return 1
+    done
+  fi
+
+  return 0
+}
+
+ontology_names=()
+for person_file in "$PERSONS_DIR"/*.md; do
+  [ -f "$person_file" ] || continue
+  while IFS= read -r person_name; do
+    person_norm="$(normalize_name "$person_name")"
+    [ -n "$person_norm" ] || continue
+    ontology_names+=("$person_norm")
+  done < <(extract_person_names "$person_file")
+done
+
+if [ "${#ontology_names[@]}" -eq 0 ]; then
+  echo "ERROR: ontology person directory has no canonical_name or aliases to compare: companies/$SLUG/ontology/entities/person" >&2
+  exit 2
+fi
+
+drift_count=0
+while IFS= read -r team_name; do
+  team_norm="$(normalize_name "$team_name")"
+  [ -n "$team_norm" ] || continue
+
+  matched=0
+  for ontology_name in "${ontology_names[@]}"; do
+    if name_matches "$team_norm" "$ontology_name"; then
+      matched=1
+      break
+    fi
+  done
+
+  if [ "$matched" -eq 0 ]; then
+    echo "DRIFT: \"$team_name\" in companies/$SLUG README team table has no matching active ontology person entity (possibly departed or misspelled)"
+    drift_count=$((drift_count + 1))
+  fi
+done < <(extract_team_names)
+
+if [ "$drift_count" -eq 0 ]; then
+  echo "ontology-readme-drift($SLUG): no drift"
+fi
+echo "ontology-readme-drift($SLUG): $drift_count drift(s)"
+
+if [ "$drift_count" -gt 0 ]; then
+  exit 1
+fi