agentsmith Agent harness 操作系统

35write_hook() {  # <hookname> <body> — managed write w/ foreign-backup; empty body removes a stale managed hook
36  local name="$1" body="$2" f="$HOOKS_DIR/$1" marker="agentsmith $1"
37  # legacy pre-rebrand marker: still recognized as "ours" so old hooks migrate in place, never written
38  local legacy_marker="universal-claude-harness $1"
39  if [ -z "$body" ]; then
40    if [ -f "$f" ] && grep -q -e "$marker" -e "$legacy_marker" "$f"; then
41      rm -f "$f"; echo "  removed managed $name (not selected)"
42    fi
43    return 0   # bare 'return' would propagate the failed test above and kill the set -e script
44  fi
45  [ -f "$f" ] && ! grep -q -e "$marker" -e "$legacy_marker" "$f" && { cp "$f" "$f.bak.$$"; echo "  backed up foreign $name → $name.bak.$$"; }
46  {
47    echo "#!/usr/bin/env bash"
48    echo "# $marker — regenerate via scripts/install-git-hooks.sh. Bypass once: --no-verify"
49    echo "set -e"
50    echo 'ROOT="$(git rev-parse --show-toplevel)"'
51    printf '%s\n' "$body"
52  } > "$f"
53  chmod +x "$f"; echo "  installed $name"

4**Unit of work:** one tracked item (an issue, a ticket, a task, a deliverable) per session by
5default. Its description is the contract. Two or three closely-related items on the same branch
6or workspace is fine when they share scope naturally. For an obvious small fix, a one-line
7description instead of a formal ticket is acceptable.
8
9**Autonomy:** proceed through **plan → do → verify → finalize → hand off** without asking for
10approval between steps. Decide routing and scope calls yourself, explain the WHY in the
11commit/PR/summary, and move on. The autonomy is in the *quantity of un-gated steps*, never in
12the *quality bar* — the principle rules and the profile's gates still apply to every step.
13
14**When the item is ambiguous:** research the right answer (official docs, reputable sources,
15the codebase/asset itself) and pick the researched path. **Do not default to "most
16conservative" or pick at random.** Note what you researched in the commit/PR/summary so the
17choice is auditable later.

52if [ ! -f "$CONF" ]; then
53  echo "${RED}No phase config at .harness/verify.conf${RESET}"
54  echo "${YELLOW}Create it (see .harness/verify.conf.example) — list the commands that prove this project is shippable.${RESET}"
55  exit 2
56fi
57
58# Parse "Label :: command" lines (skip comments/blanks).
59LABELS=(); CMDS=()
60while IFS= read -r line; do
61  [ -z "${line//[[:space:]]/}" ] && continue
62  case "$line" in \#*) continue ;; esac
63  label="${line%%::*}"; cmd="${line#*::}"
64  label="$(echo "$label" | sed 's/[[:space:]]*$//;s/^[[:space:]]*//')"
65  cmd="$(echo "$cmd" | sed 's/^[[:space:]]*//')"
66  [ -z "$cmd" ] && continue
67  if [ -n "$ONLY" ] && [[ "$label" != *"$ONLY"* ]]; then continue; fi
68  LABELS+=("$label"); CMDS+=("$cmd")
69done < "$CONF"
70
71TOTAL=${#LABELS[@]}
72if [ "$TOTAL" -eq 0 ]; then
73  echo "${YELLOW}No matching phases.${RESET}"; exit 0
74fi
75
76if $LIST_ONLY; then
77  echo "${BOLD}Configured phases:${RESET}"
78  for i in "${!LABELS[@]}"; do printf '  %d. %s :: %s\n' "$((i+1))" "${LABELS[$i]}" "${CMDS[$i]}"; done
79  exit 0
80fi
81
82cd "$ROOT_DIR"
83START=$SECONDS
84for i in "${!LABELS[@]}"; do
85  echo; echo "${BOLD}${CYAN}[$((i+1))/$TOTAL] ${LABELS[$i]}${RESET}"
86  echo "  ${CMDS[$i]}"
87  if ! bash -c "${CMDS[$i]}"; then
88    echo; echo "${BOLD}${RED}verify.sh: phase '${LABELS[$i]}' failed.${RESET}"
89    echo "${YELLOW}Run that command directly to iterate, then re-run verify.sh.${RESET}"
90    exit 1
91  fi
92done
93
94echo; echo "${BOLD}${GREEN}verify.sh: all $TOTAL phase(s) passed in $((SECONDS-START))s.${RESET}"

20# Generic high-signal secret patterns (shapes, not values).
21PATTERNS=(
22  'BEGIN [A-Z ]*PRIVATE KEY'                            # PEM private keys (RSA/OPENSSH/EC/bare/ENCRYPTED)
23  'AKIA[0-9A-Z]{16}'                                    # AWS access key id
24  'ASIA[0-9A-Z]{16}'                                    # AWS temp key id
25  'gh[pousr]_[A-Za-z0-9]{20,}'                          # GitHub tokens
26  'xox[baprs]-[A-Za-z0-9-]{10,}'                        # Slack tokens
27  'sk-[A-Za-z0-9]{20,}'                                 # OpenAI-style keys
28  'sk-ant-[A-Za-z0-9_-]{20,}'                           # Anthropic-style keys
29  'AIza[0-9A-Za-z_-]{35}'                               # Google API key
30  '(password|passwd|pwd|secret|token|api[_-]?key)["'"'"' ]*[:=][[:space:]]*["'"'"'][^"'"'"' ]{8,}'  # assigned secret literal
31)
32
33# Patterns go to a temp file so the whole scan is ONE grep pass (-f ORs every line of the file).
34# Matching per-line in a shell loop costs a subprocess per line per pattern — imperceptible on
35# Linux (~1ms spawns), minutes on Windows/Git Bash (~30ms). This script runs as a pre-commit
36# hook, and a hook slow enough to be disabled is a guardrail that isn't there (R8).
37PATTERN_FILE="$(mktemp)"
38trap 'rm -f "$PATTERN_FILE"' EXIT
39printf '%s\n' "${PATTERNS[@]}" > "$PATTERN_FILE"

73# check <label> <extended-regex> <exempt-regex|""> <hint>
74# Findings print as file:line:match, so <exempt-regex> can anchor on the match with $.
75check() {
76  local label="$1" ere="$2" exempt="$3" hint="$4" hits line
77  # -I skips binaries; -o prints just the match so exemptions test the match, not the whole
78  # line (a line holding both a safe and an unsafe value must still fail).
79  # `|| true` absorbs grep's exit-1-on-no-match, which pipefail would treat as a hard error.
80  hits="$(git grep -noIiE -- "$ere" -- . "${SCOPE[@]+"${SCOPE[@]}"}" 2>/dev/null || true)"
81  [ -n "$hits" ] || return 0
82  if [ -n "$exempt" ]; then
83    hits="$(printf '%s\n' "$hits" | grep -Evi "$exempt" || true)"
84    [ -n "$hits" ] || return 0
85  fi
86  hits="$(printf '%s\n' "$hits" | allow_filter || true)"
87  [ -n "$hits" ] || return 0
88  echo "  ✗ $label — $hint"
89  while IFS= read -r line; do
90    echo "      $line"
91  done <<<"$hits"
92  fail=1
93}

45for p in "${profiles[@]}"; do
46  name="$(basename "$p" .md)"
47  out="$WORK/$name"
48  mkdir -p "$out"
49  if ! bash "$SETUP" --profile "$name" --assemble-only --target "$out" \
50        --operator-name "Test" --tracker "Linear" >/dev/null 2>&1; then
51    bad "$name: assembly failed outright"
52    continue
53  fi
54  f="$out/CLAUDE.md"
55  if [ ! -s "$f" ]; then
56    bad "$name: produced no CLAUDE.md"
57    continue
58  fi
59
60  errs=""
61  # An unrendered token means the agent reads a literal "{{TRACKER}}" as if it were the rule.
62  # Never acceptable, in any profile.
63  leaked="$(grep -oE '\{\{[A-Z_]+\}\}' "$f" 2>/dev/null | sort -u | tr '\n' ' ')"
64  [ -n "$leaked" ] && errs="$errs unrendered:[$leaked]"
65  # [TODO: set X] means "the assembler had no value for X". Two very different cases:
66  #   - X is flag-driven (below) → setup COULD have filled it and didn't. That's a bug: the agent
67  #     gets a blank where a rule should be. This is the regression that would fire if a
68  #     placeholder were renamed in core/ but not in fill_placeholders().
69  #   - X has no flag (BRAND_PALETTE, BRAND_FONT) → the TODO IS the product. It's the documented
70  #     handoff to the human (INSTALL.md's placeholder table; setup's own "resolve any [TODO]"
71  #     next-step). Flagging it would make this suite red by design and train people to ignore it.
72  for tok in OPERATOR_NAME OPERATOR_ROLE OPERATOR_BIO TRACKER TRACKER_POLICY; do
73    grep -qF "[TODO: set $tok]" "$f" 2>/dev/null && errs="$errs unfilled-flag-token:[$tok]"
74  done
75
76  if [ -n "$errs" ]; then
77    bad "$name: assembled with leaks —$errs"
78  else
79    ok "$name: assembles clean (no unrendered tokens, no TODO blanks)"
80  fi