fix(tui): establish 2-tier readiness model, modal/hint token separation, and adapter modal contracts

- Separate TUI dialog detection into blocking gate layer (_MAM_MODAL_TOKENS) and recovery layer (_MAM_HINT_TOKENS) to eliminate idle dialog starvation
- Add modal_tokens contract across agent adapters (claude, cline) with dynamic scope facts resolution in send_keys_safe
- Implement fail-closed resolution for ambiguous same-kind panes (R-1) and enforce WORKSPACE_ROOT export (R-2)
- Preserve session on TUI readiness timeout when PID is alive for diagnostics
- Add comprehensive test suite in tests/test_c1_tui_readiness.py and adapter contract assertions in test_a4
This commit is contained in:
2026-08-28 15:21:10 +09:00
parent 4a5a093986
commit 17edf90676
9 changed files with 868 additions and 42 deletions
+156 -33
View File
@@ -24,6 +24,7 @@ fi
SKILL_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
export PYTHONPATH="$SKILL_DIR:${PYTHONPATH:-}"
WORKSPACE_ROOT="${WORKSPACE_ROOT:-$(cd "$SKILL_DIR/../.." && pwd)}"
export WORKSPACE_ROOT
AGENT_SESSIONS_YAML="${AGENT_SESSIONS_YAML:-$WORKSPACE_ROOT/.mam/agent-sessions.yaml}"
_sanitize_herdr_agent_name() {
@@ -43,12 +44,14 @@ _sanitize_herdr_agent_name() {
h=$(printf '%s' "$s" | sha1sum 2>/dev/null | awk '{print $1}' | cut -c 1-8)
elif command -v openssl >/dev/null 2>&1; then
h=$(printf '%s' "$s" | openssl sha1 2>/dev/null | awk '{print $NF}' | cut -c 1-8)
else
h=$(python3 -c "import hashlib,sys; print(hashlib.sha1(sys.argv[1].encode()).hexdigest()[:8])" "$s" 2>/dev/null || echo "00000000")
fi
s="${s:0:23}-${h}"
if [ -n "$h" ]; then
s="${s:0:23}-$h"
else
s="${s:0:32}"
fi
fi
printf '%s\n' "${s:0:32}"
printf '%s\n' "$s"
}
# Add common Homebrew and local binary paths to PATH to ensure they are available in non-interactive shells
@@ -59,7 +62,9 @@ for dir in /home/linuxbrew/.linuxbrew/bin /home/linuxbrew/.linuxbrew/sbin "$HOME
done
# Central TUI dialog and readiness validation tokens (OP-6)
_MAM_DIALOG_TOKENS='Do you trust the files|Yes, proceed|No, exit|Allow this|Press Enter to continue|browser to authenticate|Use arrow keys|Esc to cancel|Resuming the full session|Resume from summary'
_MAM_MODAL_TOKENS='Do you trust the files|Yes, proceed|No, exit|Allow this|browser to authenticate|Resuming the full session|Resume from summary|Try the new fullscreen renderer\?'
_MAM_HINT_TOKENS='Use arrow keys|Esc to cancel|Press Enter to continue'
_MAM_DIALOG_TOKENS="${_MAM_MODAL_TOKENS}|${_MAM_HINT_TOKENS}"
_MAM_READY_TOKENS_CLAUDE='Anthropic|Assistant|Chat|Welcome|Claude Code|Opus|Sonnet|Haiku'
# Workspace-relative defaults with environment overrides (Phase Z)
@@ -348,17 +353,18 @@ _resolve_herdr_pane_id() {
local ws_flag=()
[ -n "$target_ws" ] && ws_flag=(--workspace "$target_ws")
pid=$(_real_herdr pane list "${ws_flag[@]+"${ws_flag[@]}"}" 2>/dev/null \
| TARGET_NAME="$target" TARGET_SAN="$sat" TARGET_WS="$target_ws" python3 -c "
| TARGET_NAME="$target" TARGET_SAN="$sat" TARGET_WS="$target_ws" WORKSPACE_ROOT="${WORKSPACE_ROOT:-$PWD}" python3 -c "
import sys, json, os
tn = os.environ.get('TARGET_NAME', '')
tsa = os.environ.get('TARGET_SAN', '')
tws = os.environ.get('TARGET_WS', '')
ws_root = os.environ.get('WORKSPACE_ROOT', os.getcwd())
try:
panes = json.load(sys.stdin).get('result', {}).get('panes', [])
# Client-side filter in case an older herdr ignores --workspace.
if tws:
panes = [p for p in panes if p.get('workspace_id') == tws]
# ISSUE-2: exact match only.
# ISSUE-2: exact match first, then agent-sessions.yaml lookup fallback
for key in ('label', 'name', 'agent'):
for p in panes:
v = p.get(key)
@@ -367,6 +373,27 @@ try:
if pid:
print(pid)
sys.exit(0)
agent_kind = None
yaml_path = os.path.join(ws_root, '.mam', 'agent-sessions.yaml')
if os.path.isfile(yaml_path):
try:
import yaml
with open(yaml_path) as f:
ydata = yaml.safe_load(f) or {}
for s in ydata.get('herdr_sessions', []):
if s.get('name') == tn or s.get('name') == tsa:
agent_kind = s.get('pane', {}).get('cmd') or ''
break
except Exception as e:
sys.stderr.write(f'warning: yaml load failed: {e}\n')
if agent_kind:
matching = [p.get('pane_id') for p in panes if p.get('agent') == agent_kind and p.get('pane_id')]
if len(matching) == 1:
print(matching[0])
sys.exit(0)
elif len(matching) > 1:
# R-1: Ambiguous resolution with multiple same-kind panes — fail closed
sys.exit(1)
except Exception:
pass
sys.exit(1)
@@ -731,13 +758,13 @@ except Exception:
else
res=$(eval "_real_herdr agent start \"$agent_name\" --kind \"$kind\" --pane \"$target_pane\"" 2>&1 || true)
fi
# Fatal CLI errors abort immediately. agent_not_ready is herdr's
# documented "process is up, blocked on a dialog" status — continue
# to wait_for_tui_ready. Do NOT treat "timed out waiting for agent
# startup" as success: herdr returns that for a dead process too.
if echo "$res" | grep -qiE "^usage:|unknown option|unknown flag|missing required|invalid_agent_name|^error:"; then
break
fi
# Fatal CLI errors abort immediately. agent_not_ready is the
# documented "process is up, blocked on a dialog" status — continue
# to wait_for_tui_ready. Do NOT treat "timed out waiting for agent
# startup" as success: herdr returns that for a dead process too.
if echo "$res" | grep -qE "agent_started|agent_not_ready"; then
success=1
break
@@ -1793,41 +1820,117 @@ start_watchdog() {
# wait_for_tui_ready <session_name> <agent>
# Waits up to 30 seconds for the agent's TUI to render its welcome screen.
# Returns 0 on readiness, 1 on timeout, 2 on unobservable/headless.
wait_for_tui_ready() {
local sess="$1" agent="$2"
# Self-contained resolution: works whether or not caller evaluated facts (C1-(3)).
local tokens="${MAM_READY_TOKENS:-}"
if [ -z "$tokens" ]; then
local strong_tokens="${MAM_STRONG_READY_TOKENS:-}"
local weak_tokens="${MAM_WEAK_READY_TOKENS:-}"
local placeholder="${MAM_INPUT_PLACEHOLDER:-}"
local prompt="${MAM_INPUT_PROMPT:-}"
local rule_pattern="${MAM_INPUT_RULE_PATTERN:-}"
local modal_tokens="${MAM_MODAL_TOKENS:-}"
# Self-contained resolution: fetch facts in 1 call if not already in environment
if [ -z "$strong_tokens" ] && [ -z "$weak_tokens" ] && [ -z "${MAM_READY_TOKENS:-}" ]; then
local _facts=""
_facts="$("$(_delegate_py_bin)" -m lib_py.agents facts "$agent" 2>/dev/null)" || _facts=""
tokens="$(printf '%s\n' "$_facts" | sed -n 's/^MAM_READY_TOKENS=//p')"
[ -n "$tokens" ] && eval "tokens=$tokens"
if [ -n "$_facts" ]; then
eval "$_facts"
strong_tokens="${MAM_STRONG_READY_TOKENS:-}"
weak_tokens="${MAM_WEAK_READY_TOKENS:-}"
placeholder="${MAM_INPUT_PLACEHOLDER:-}"
prompt="${MAM_INPUT_PROMPT:-}"
rule_pattern="${MAM_INPUT_RULE_PATTERN:-}"
modal_tokens="${MAM_MODAL_TOKENS:-}"
fi
fi
if [ -z "$tokens" ]; then
# Fallback if strong_tokens is empty but MAM_READY_TOKENS is set
if [ -z "$strong_tokens" ] && [ -n "${MAM_READY_TOKENS:-}" ]; then
strong_tokens="$MAM_READY_TOKENS"
fi
if [ -z "$strong_tokens" ] && [ -z "$weak_tokens" ] && [ -z "$placeholder" ]; then
echo "wait_for_tui_ready: no ready tokens for agent '$agent'" >&2
return 1
fi
local i
local empty_streak=0
local empty_giveup="${MAM_READY_EMPTY_GIVEUP:-5}"
local modal_pat="$_MAM_MODAL_TOKENS"
[ -n "$modal_tokens" ] && modal_pat="${modal_pat}|${modal_tokens}"
local i content tail_lines
for i in {1..30}; do
if [ "$agent" = "claude" ]; then
handle_startup_dialogs "$sess" 1 || true
fi
if _pane_dialog_open "$sess"; then
if printf '%s\n' "$(_pane_tail "$sess" 5)" | grep -q 'Press Enter to continue'; then
_sks_herdr send-keys -t "$sess" Enter || true
sleep 1
# 1. Single capture per iteration (T-3)
content=$(_pane_capture "$sess")
# Consecutive blank capture check (C-6)
if [ -z "$content" ]; then
empty_streak=$((empty_streak + 1))
if [ "$empty_streak" -ge "$empty_giveup" ]; then
echo "wait_for_tui_ready: unobservable/headless pane for '$sess' (consecutive empty captures: $empty_streak)" >&2
return 2
fi
sleep 1
continue
fi
local content
content=$(_sks_herdr capture-pane -p -t "$sess" 2>/dev/null || echo "")
if [ -n "$content" ]; then
if echo "$content" | grep -E -q "$tokens" 2>/dev/null; then
echo "$agent TUI detected ready."
return 0
empty_streak=0
# 2. Check modal dialogs FIRST (F-2): if a blocking modal signature is present in the tail lines,
# handle it and do NOT declare ready on this iteration.
tail_lines=$(printf '%s\n' "$content" | grep -v '^[[:space:]]*$' | tail -n 20)
if printf '%s\n' "$tail_lines" | grep -E -q "$modal_pat" 2>/dev/null; then
handle_startup_dialogs "$sess" 1 || true
sleep 1
continue
fi
# 3. Tier 2 Readiness Evaluation (T-2b, T-4: Ready check first)
local s_matched=false
local w_matched=false
local c_matched=false
# S-layer: Strong ready token match
if [ -n "$strong_tokens" ]; then
if echo "$content" | grep -E -q "$strong_tokens" 2>/dev/null; then
s_matched=true
fi
fi
# S-layer: Placeholder match (must be >= 8 characters)
if [ -n "$placeholder" ] && [ "${#placeholder}" -ge 8 ]; then
if echo "$content" | grep -F -q "$placeholder" 2>/dev/null; then
s_matched=true
fi
fi
# W-layer: Weak ready token match
if [ -n "$weak_tokens" ]; then
if echo "$content" | grep -E -q "$weak_tokens" 2>/dev/null; then
w_matched=true
fi
fi
# C-layer: Corroboration (Prompt in last 5 non-blank lines OR rule pattern match)
if [ -n "$prompt" ]; then
if printf '%s\n' "$content" | grep -v '^[[:space:]]*$' | tail -n 5 | grep -F -q "$prompt" 2>/dev/null; then
c_matched=true
fi
fi
if [ -n "$rule_pattern" ]; then
if echo "$content" | grep -E -q "$rule_pattern" 2>/dev/null; then
c_matched=true
fi
fi
# Decision: S OR (W AND C)
if [ "$s_matched" = true ] || { [ "$w_matched" = true ] && [ "$c_matched" = true ]; }; then
echo "$agent TUI detected ready."
return 0
fi
# 4. Startup dialog & hint resolution if not ready
handle_startup_dialogs "$sess" 1 || true
sleep 1
done
echo "⚠️ TUI readiness check timed out for '$sess'." >&2
@@ -1920,11 +2023,14 @@ _pane_quiescent() {
}
# _pane_dialog_open <sess> — focus-stealing modal signatures (trust /
# permission / OAuth / list-selection), checked in the bottom 20 pane lines
# permission / OAuth / list-selection), checked in the bottom 8 pane lines
# only (dialogs render near the input area; conversation text above must not
# trigger this). Tokens must NOT appear on normal idle prompt screens.
_pane_dialog_open() {
_pane_tail "$1" 20 | grep -Eq "$_MAM_DIALOG_TOKENS"
local extra="${MAM_MODAL_TOKENS:-}"
local pat="$_MAM_MODAL_TOKENS"
[ -n "$extra" ] && pat="${pat}|${extra}"
_pane_tail "$1" "${MAM_DIALOG_TAIL_LINES:-8}" | grep -Eq "$pat"
}
# send_keys_safe <sess> <text> [job_id]
@@ -1939,6 +2045,23 @@ _pane_dialog_open() {
send_keys_safe() {
local sess="$1" text="$2" job_id="${3:-adhoc}"
local pre_submit deadline try
local MAM_MODAL_TOKENS="${MAM_MODAL_TOKENS:-}"
local MAM_AGENT_NAME MAM_OWN_KEY MAM_INPUT_PROMPT MAM_INPUT_PLACEHOLDER MAM_INPUT_RULE_PATTERN
local MAM_READY_TOKENS MAM_STRONG_READY_TOKENS MAM_WEAK_READY_TOKENS MAM_EXIT_KEY MAM_DELEGATE_AGENT_KEY
if [ -z "${MAM_MODAL_TOKENS:-}" ]; then
local _sks_agent=""
case "$sess" in
*-claude|*-claude-[0-9]*|claude|claude-[0-9]*) _sks_agent="claude" ;;
*-agy|*-agy-[0-9]*|agy|agy-[0-9]*) _sks_agent="agy" ;;
*-hermes|*-hermes-[0-9]*|hermes|hermes-[0-9]*) _sks_agent="hermes" ;;
*-cline|*-cline-[0-9]*|cline|cline-[0-9]*) _sks_agent="cline" ;;
*-grok|*-grok-[0-9]*|grok|grok-[0-9]*) _sks_agent="grok" ;;
esac
if [ -n "$_sks_agent" ]; then
eval "$("$(_delegate_py_bin)" -m lib_py.agents facts "$_sks_agent" 2>/dev/null || true)"
fi
fi
local _q_rc=0
_pane_quiescent "$sess" "${SKS_QUIESCENT_TRIES:-20}" "${SKS_QUIESCENT_INTERVAL:-0.5}" || _q_rc=$?
+3
View File
@@ -23,6 +23,9 @@ def main():
print(f"MAM_INPUT_PLACEHOLDER={q(adapter.input_placeholder or '')}")
print(f"MAM_INPUT_RULE_PATTERN={q(adapter.input_rule_pattern or '')}")
print(f"MAM_READY_TOKENS={q(adapter.ready_tokens)}")
print(f"MAM_STRONG_READY_TOKENS={q(adapter.strong_ready_tokens)}")
print(f"MAM_WEAK_READY_TOKENS={q(adapter.weak_ready_tokens)}")
print(f"MAM_MODAL_TOKENS={q(adapter.modal_tokens or '')}")
print(f"MAM_EXIT_KEY={q(adapter.exit_key)}")
print(f"MAM_DELEGATE_AGENT_KEY={q(adapter.delegate_agent_key)}")
@@ -40,6 +40,10 @@ class ClaudeAgentAdapter(BaseAgentAdapter):
def input_rule_pattern(self) -> str:
return '{10,}'
@property
def modal_tokens(self) -> Optional[str]:
return 'Try the new fullscreen renderer\\?'
def artifact_path(self, uuid: str, ctx: DiscoveryContext) -> str:
return f"{ctx.claude_dir}/{ctx.ws_key}/{uuid}.jsonl"
@@ -16,6 +16,18 @@ class ClineAgentAdapter(BaseAgentAdapter):
def ready_tokens(self) -> str:
return 'Cline|history|Chat|What can I do|slash commands'
@property
def strong_ready_tokens(self) -> str:
return 'Cline|What can I do|slash commands'
@property
def weak_ready_tokens(self) -> str:
return 'history|Chat'
@property
def modal_tokens(self) -> Optional[str]:
return 'Select API Provider|Enter API Key|Select an API provider|API Provider|Cline API key'
@property
def exit_key(self) -> str:
return '/exit'
+12
View File
@@ -73,6 +73,18 @@ class BaseAgentAdapter:
def input_rule_pattern(self) -> Optional[str]:
return None
@property
def modal_tokens(self) -> Optional[str]:
return None
@property
def strong_ready_tokens(self) -> str:
return self.ready_tokens
@property
def weak_ready_tokens(self) -> str:
return ''
def derive_session_name(self, slug: str, role: str = "creator") -> str:
r = role.lower()
return f"{slug}-{r}-{self.name}"
@@ -225,7 +225,7 @@ spawn
cleanup_herdr_on_error() {
local exit_code=$?
if [ $exit_code -ne 0 ]; then
echo "⚠️ Error occurred during initialization. Rolling back and killing herdr session '$SESSION_NAME'..." >&2
echo "⚠️ Error occurred during initialization (exit code $exit_code). Rolling back and killing herdr session '$SESSION_NAME'..." >&2
_herdr kill-session -t "$SESSION_NAME" 2>/dev/null || true
fi
}
@@ -237,11 +237,48 @@ if [ -z "$HERDR_SERVER_OPT" ]; then
fi
# TUI 준비 대기
if ! wait_for_tui_ready "$SESSION_NAME" "$AGENT"; then
echo "ERROR: agent TUI never became ready — aborting (rollback via trap)" >&2
exit 1
TUI_READY_RC=0
wait_for_tui_ready "$SESSION_NAME" "$AGENT" || TUI_READY_RC=$?
LAST_VISIBLE_STATUS_OVERRIDE=""
if [ "$TUI_READY_RC" -ne 0 ]; then
# 1. Diagnostic dump
DIAG_DIR="${WORKSPACE:-$PWD}/.mam/diagnostics"
mkdir -p "$DIAG_DIR" 2>/dev/null || true
DIAG_FILE="$DIAG_DIR/${SESSION_NAME}-$(date +%s).txt"
{
echo "=== MAM Diagnostic Dump ==="
echo "Date: $(date -u +'%Y-%m-%dT%H:%M:%SZ')"
echo "Session: $SESSION_NAME"
echo "Agent: $AGENT"
echo "ReturnCode: $TUI_READY_RC"
echo "--- Pane Raw Capture ---"
_pane_capture "$SESSION_NAME" 2>/dev/null || true
} > "$DIAG_FILE" 2>/dev/null || true
if [ ! -f "$DIAG_FILE" ]; then
echo "WARNING: Failed to write diagnostic dump to $DIAG_FILE" >&2
fi
# 2. Check pane PID liveness
PANE_PID_CHECK=$(_herdr list-panes -t "$SESSION_NAME" -F '#{pane_pid}' 2>/dev/null || echo "")
ALIVE=false
if [ -n "$PANE_PID_CHECK" ] && [[ "$PANE_PID_CHECK" =~ ^[0-9]+$ ]]; then
if kill -0 "$PANE_PID_CHECK" 2>/dev/null; then
ALIVE=true
fi
fi
if [ "$ALIVE" = true ] && [ "${MAM_KILL_ON_READY_TIMEOUT:-0}" != "1" ]; then
echo "WARNING: agent TUI never became ready within timeout, but process (PID $PANE_PID_CHECK) is alive. Preserving session for diagnosis." >&2
LAST_VISIBLE_STATUS_OVERRIDE="tui-ready-timeout"
SUBMIT_JOB_PROMPT="" # disable instruction injection on unready session
else
echo "ERROR: agent TUI never became ready — aborting (rollback via trap)" >&2
exit 42
fi
fi
if [ "$AGENT" = "claude" ]; then
if [ "$AGENT" = "claude" ] && [ -z "$LAST_VISIBLE_STATUS_OVERRIDE" ]; then
handle_startup_dialogs "$SESSION_NAME" 15
fi
@@ -257,8 +294,8 @@ NOW_ISO=$(date -u +'%Y-%m-%dT%H:%M:%SZ')
# herdr server shim (HERDR_SESSION_NAME picked up by the lib.sh shim).
START_CMD="HERDR_SESSION_NAME=${HERDR_SESSION_NAME:-default} herdr new-session -d -s \"$SESSION_NAME\" -x 140 -y 40 -c \"$WORKSPACE\" \"$CMD_FULL\""
# If --onboard is specified, automatically build the onboarding prompt
if [ "$ONBOARD" = "1" ] && [ -z "$SUBMIT_JOB_PROMPT" ]; then
# If --onboard is specified, automatically build the onboarding prompt (if session is ready)
if [ "$ONBOARD" = "1" ] && [ -z "$SUBMIT_JOB_PROMPT" ] && [ -z "$LAST_VISIBLE_STATUS_OVERRIDE" ]; then
SUBMIT_JOB_PROMPT="You are a newly spawned $ROLE Team Leader agent in this workspace. Without making any non-standard pre-preparations or modifying files directly, proceed immediately to align yourself with the project context by performing the following tasks:
1. Read the project documentation at README.md and the multi-agent protocol guidelines at .agents/MULTI_AGENT_RULES.md to understand the design rules.
2. Run 'git status' and 'git diff' to analyze the current modifications and active work in the repository.
@@ -268,6 +305,9 @@ fi
# agent-sessions.yaml 에 append
DELEGATE_JOB_ID=""
if [ -n "$LAST_VISIBLE_STATUS_OVERRIDE" ]; then
SUBMIT_JOB_PROMPT=""
fi
if [ -n "$SUBMIT_JOB_PROMPT" ]; then
delegate_agent="${MAM_DELEGATE_AGENT_KEY:-}"
if [ -z "$delegate_agent" ]; then
@@ -306,6 +346,7 @@ atomic_dump_yaml "$AGENT_SESSIONS_YAML" \
HERDR_SESSION_NAME="${HERDR_SESSION_NAME:-default}" \
MAM_WS_LABEL="$MAM_WS_LABEL" \
SESSION_UUID="$SESSION_UUID" \
LAST_VISIBLE_STATUS_OVERRIDE="${LAST_VISIBLE_STATUS_OVERRIDE:-}" \
DELEGATE_JOB_ID="$DELEGATE_JOB_ID" ROLE="$ROLE" <<'PYEOF'
name = os.environ['SESSION_NAME']
agent = os.environ['AGENT']
@@ -389,6 +430,10 @@ elif agent == 'grok':
entry['session_id_verified'] = False
entry['last_visible_status'] = "assigned (awaiting first message)" if assigned else "unverified"
last_vis_override = os.environ.get('LAST_VISIBLE_STATUS_OVERRIDE', '')
if last_vis_override:
entry['last_visible_status'] = last_vis_override
sessions.append(entry)
snap = d.setdefault('snapshot', {})