test: cover claude and codex acp bind docker smoke

2026-04-21 22:21:33 +00:00 · 2026-04-07 06:06:13 +01:00
parent c2cd1aed5d
commit ce1d2c1004
4 changed files with 228 additions and 111 deletions
--- a/docs/help/testing.md
+++ b/docs/help/testing.md
@@ -305,12 +305,14 @@ Notes:
  - `pnpm test:live src/gateway/gateway-acp-bind.live.test.ts`
  - `OPENCLAW_LIVE_ACP_BIND=1`
 - Defaults:
-  - ACP agent: `claude`
+  - ACP agents in Docker: `claude,codex`
+  - ACP agent for direct `pnpm test:live ...`: `claude`
  - Synthetic channel: Slack DM-style conversation context
  - ACP backend: `acpx`
 - Overrides:
  - `OPENCLAW_LIVE_ACP_BIND_AGENT=claude`
  - `OPENCLAW_LIVE_ACP_BIND_AGENT=codex`
+  - `OPENCLAW_LIVE_ACP_BIND_AGENTS=claude,codex`
  - `OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND='npx -y @agentclientprotocol/claude-agent-acp@<version>'`
 - Notes:
  - This lane uses the gateway `chat.send` surface with admin-only synthetic originating-route fields so tests can attach message-channel context without pretending to deliver externally.
@@ -330,9 +332,18 @@ Docker recipe:
 pnpm test:docker:live-acp-bind
 ```

+Single-agent Docker recipes:
+
+```bash
+pnpm test:docker:live-acp-bind:claude
+pnpm test:docker:live-acp-bind:codex
+```
+
 Docker notes:

 - The Docker runner lives at `scripts/test-live-acp-bind-docker.sh`.
+- By default, it runs the ACP bind smoke against both supported live CLI agents in sequence: `claude`, then `codex`.
+- Use `OPENCLAW_LIVE_ACP_BIND_AGENTS=claude` or `OPENCLAW_LIVE_ACP_BIND_AGENTS=codex` to narrow the matrix.
 - It sources `~/.profile`, stages the matching CLI auth material into the container, installs `acpx` into a writable npm prefix, then installs the requested live CLI (`@anthropic-ai/claude-code` or `@openai/codex`) if missing.
 - Inside Docker, the runner sets `OPENCLAW_LIVE_ACP_BIND_ACPX_COMMAND=$HOME/.npm-global/bin/acpx` so acpx keeps provider env vars from the sourced profile available to the child harness CLI.

--- a/package.json
+++ b/package.json
@@ -1169,6 +1169,8 @@
    "test:docker:doctor-switch": "bash scripts/e2e/doctor-install-switch-docker.sh",
    "test:docker:gateway-network": "bash scripts/e2e/gateway-network-docker.sh",
    "test:docker:live-acp-bind": "bash scripts/test-live-acp-bind-docker.sh",
+    "test:docker:live-acp-bind:claude": "OPENCLAW_LIVE_ACP_BIND_AGENT=claude bash scripts/test-live-acp-bind-docker.sh",
+    "test:docker:live-acp-bind:codex": "OPENCLAW_LIVE_ACP_BIND_AGENT=codex bash scripts/test-live-acp-bind-docker.sh",
    "test:docker:live-build": "bash scripts/test-live-build-docker.sh",
    "test:docker:live-cli-backend": "bash scripts/test-live-cli-backend-docker.sh",
    "test:docker:live-gateway": "bash scripts/test-live-gateway-models-docker.sh",
--- a/scripts/test-live-acp-bind-docker.sh
+++ b/scripts/test-live-acp-bind-docker.sh
@@ -9,24 +9,26 @@ CONFIG_DIR="${OPENCLAW_CONFIG_DIR:-$HOME/.openclaw}"
 WORKSPACE_DIR="${OPENCLAW_WORKSPACE_DIR:-$HOME/.openclaw/workspace}"
 PROFILE_FILE="${OPENCLAW_PROFILE_FILE:-$HOME/.profile}"
 CLI_TOOLS_DIR="${OPENCLAW_DOCKER_CLI_TOOLS_DIR:-$HOME/.cache/openclaw/docker-cli-tools}"
-ACP_AGENT="${OPENCLAW_LIVE_ACP_BIND_AGENT:-claude}"
+ACP_AGENT_LIST_RAW="${OPENCLAW_LIVE_ACP_BIND_AGENTS:-${OPENCLAW_LIVE_ACP_BIND_AGENT:-claude,codex}}"

-case "$ACP_AGENT" in
-  claude)
-    AUTH_PROVIDER="claude-cli"
-    CLI_PACKAGE="@anthropic-ai/claude-code"
-    CLI_BIN="claude"
-    ;;
-  codex)
-    AUTH_PROVIDER="codex-cli"
-    CLI_PACKAGE="@openai/codex"
-    CLI_BIN="codex"
-    ;;
-  *)
-    echo "Unsupported OPENCLAW_LIVE_ACP_BIND_AGENT: $ACP_AGENT (expected claude or codex)" >&2
-    exit 1
-    ;;
-esac
+openclaw_live_acp_bind_resolve_auth_provider() {
+  case "${1:-}" in
+    claude) printf '%s\n' "claude-cli" ;;
+    codex) printf '%s\n' "codex-cli" ;;
+    *)
+      echo "Unsupported OPENCLAW_LIVE_ACP_BIND agent: ${1:-} (expected claude or codex)" >&2
+      return 1
+      ;;
+  esac
+}
+
+openclaw_live_acp_bind_resolve_agent_command() {
+  case "${1:-}" in
+    claude) printf '%s' "${OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND_CLAUDE:-${OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND:-}}" ;;
+    codex) printf '%s' "${OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND_CODEX:-${OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND:-}}" ;;
+    *) return 1 ;;
+  esac
+}

 mkdir -p "$CLI_TOOLS_DIR"

@@ -35,54 +37,6 @@ if [[ -f "$PROFILE_FILE" ]]; then
  PROFILE_MOUNT=(-v "$PROFILE_FILE":/home/node/.profile:ro)
 fi

-AUTH_DIRS=()
-AUTH_FILES=()
-if [[ -n "${OPENCLAW_DOCKER_AUTH_DIRS:-}" ]]; then
-  while IFS= read -r auth_dir; do
-    [[ -n "$auth_dir" ]] || continue
-    AUTH_DIRS+=("$auth_dir")
-  done < <(openclaw_live_collect_auth_dirs)
-  while IFS= read -r auth_file; do
-    [[ -n "$auth_file" ]] || continue
-    AUTH_FILES+=("$auth_file")
-  done < <(openclaw_live_collect_auth_files)
-else
-  while IFS= read -r auth_dir; do
-    [[ -n "$auth_dir" ]] || continue
-    AUTH_DIRS+=("$auth_dir")
-  done < <(openclaw_live_collect_auth_dirs_from_csv "$AUTH_PROVIDER")
-  while IFS= read -r auth_file; do
-    [[ -n "$auth_file" ]] || continue
-    AUTH_FILES+=("$auth_file")
-  done < <(openclaw_live_collect_auth_files_from_csv "$AUTH_PROVIDER")
-fi
-AUTH_DIRS_CSV=""
-if ((${#AUTH_DIRS[@]} > 0)); then
-  AUTH_DIRS_CSV="$(openclaw_live_join_csv "${AUTH_DIRS[@]}")"
-fi
-AUTH_FILES_CSV=""
-if ((${#AUTH_FILES[@]} > 0)); then
-  AUTH_FILES_CSV="$(openclaw_live_join_csv "${AUTH_FILES[@]}")"
-fi
-
-EXTERNAL_AUTH_MOUNTS=()
-if ((${#AUTH_DIRS[@]} > 0)); then
-  for auth_dir in "${AUTH_DIRS[@]}"; do
-    host_path="$HOME/$auth_dir"
-    if [[ -d "$host_path" ]]; then
-      EXTERNAL_AUTH_MOUNTS+=(-v "$host_path":/host-auth/"$auth_dir":ro)
-    fi
-  done
-fi
-if ((${#AUTH_FILES[@]} > 0)); then
-  for auth_file in "${AUTH_FILES[@]}"; do
-    host_path="$HOME/$auth_file"
-    if [[ -f "$host_path" ]]; then
-      EXTERNAL_AUTH_MOUNTS+=(-v "$host_path":/host-auth-files/"$auth_file":ro)
-    fi
-  done
-fi
-
 read -r -d '' LIVE_TEST_CMD <<'EOF' || true
 set -euo pipefail
 [ -f "$HOME/.profile" ] && source "$HOME/.profile" || true
@@ -152,6 +106,13 @@ cleanup() {
 trap cleanup EXIT
 source /src/scripts/lib/live-docker-stage.sh
 openclaw_live_stage_source_tree "$tmp_dir"
+# Use a writable node_modules overlay in the temp repo. Vite writes bundled
+# config artifacts under the nearest node_modules/.vite-temp path, and the
+# build-stage /app/node_modules tree is root-owned in this Docker lane.
+mkdir -p "$tmp_dir/node_modules"
+cp -aRs /app/node_modules/. "$tmp_dir/node_modules"
+rm -rf "$tmp_dir/node_modules/.vite-temp"
+mkdir -p "$tmp_dir/node_modules/.vite-temp"
 openclaw_live_link_runtime_tree "$tmp_dir"
 openclaw_live_stage_state_dir "$tmp_dir/.openclaw-state"
 openclaw_live_prepare_staged_config
@@ -163,34 +124,102 @@ EOF
 echo "==> Build live-test image: $LIVE_IMAGE_NAME (target=build)"
 docker build --target build -t "$LIVE_IMAGE_NAME" -f "$ROOT_DIR/Dockerfile" "$ROOT_DIR"

-echo "==> Run ACP bind live test in Docker"
-echo "==> Agent: $ACP_AGENT"
-echo "==> Auth dirs: ${AUTH_DIRS_CSV:-none}"
-echo "==> Auth files: ${AUTH_FILES_CSV:-none}"
-docker run --rm -t \
-  -u node \
-  --entrypoint bash \
-  -e ANTHROPIC_API_KEY \
-  -e ANTHROPIC_API_KEY_OLD \
-  -e OPENCLAW_LIVE_ACP_BIND_ANTHROPIC_API_KEY="${ANTHROPIC_API_KEY:-}" \
-  -e OPENCLAW_LIVE_ACP_BIND_ANTHROPIC_API_KEY_OLD="${ANTHROPIC_API_KEY_OLD:-}" \
-  -e OPENAI_API_KEY \
-  -e COREPACK_ENABLE_DOWNLOAD_PROMPT=0 \
-  -e HOME=/home/node \
-  -e NODE_OPTIONS=--disable-warning=ExperimentalWarning \
-  -e OPENCLAW_SKIP_CHANNELS=1 \
-  -e OPENCLAW_VITEST_FS_MODULE_CACHE=0 \
-  -e OPENCLAW_DOCKER_AUTH_DIRS_RESOLVED="$AUTH_DIRS_CSV" \
-  -e OPENCLAW_DOCKER_AUTH_FILES_RESOLVED="$AUTH_FILES_CSV" \
-  -e OPENCLAW_LIVE_TEST=1 \
-  -e OPENCLAW_LIVE_ACP_BIND=1 \
-  -e OPENCLAW_LIVE_ACP_BIND_AGENT="$ACP_AGENT" \
-  -e OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND="${OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND:-}" \
-  -v "$ROOT_DIR":/src:ro \
-  -v "$CONFIG_DIR":/home/node/.openclaw \
-  -v "$WORKSPACE_DIR":/home/node/.openclaw/workspace \
-  -v "$CLI_TOOLS_DIR":/home/node/.npm-global \
-  "${EXTERNAL_AUTH_MOUNTS[@]}" \
-  "${PROFILE_MOUNT[@]}" \
-  "$LIVE_IMAGE_NAME" \
-  -lc "$LIVE_TEST_CMD"
+IFS=',' read -r -a ACP_AGENT_TOKENS <<<"$ACP_AGENT_LIST_RAW"
+ACP_AGENTS=()
+for token in "${ACP_AGENT_TOKENS[@]}"; do
+  agent="$(openclaw_live_trim "$token")"
+  [[ -n "$agent" ]] || continue
+  openclaw_live_acp_bind_resolve_auth_provider "$agent" >/dev/null
+  ACP_AGENTS+=("$agent")
+done
+
+if ((${#ACP_AGENTS[@]} == 0)); then
+  echo "No ACP bind agents selected. Use OPENCLAW_LIVE_ACP_BIND_AGENTS=claude,codex." >&2
+  exit 1
+fi
+
+for ACP_AGENT in "${ACP_AGENTS[@]}"; do
+  AUTH_PROVIDER="$(openclaw_live_acp_bind_resolve_auth_provider "$ACP_AGENT")"
+  AGENT_COMMAND="$(openclaw_live_acp_bind_resolve_agent_command "$ACP_AGENT")"
+
+  AUTH_DIRS=()
+  AUTH_FILES=()
+  if [[ -n "${OPENCLAW_DOCKER_AUTH_DIRS:-}" ]]; then
+    while IFS= read -r auth_dir; do
+      [[ -n "$auth_dir" ]] || continue
+      AUTH_DIRS+=("$auth_dir")
+    done < <(openclaw_live_collect_auth_dirs)
+    while IFS= read -r auth_file; do
+      [[ -n "$auth_file" ]] || continue
+      AUTH_FILES+=("$auth_file")
+    done < <(openclaw_live_collect_auth_files)
+  else
+    while IFS= read -r auth_dir; do
+      [[ -n "$auth_dir" ]] || continue
+      AUTH_DIRS+=("$auth_dir")
+    done < <(openclaw_live_collect_auth_dirs_from_csv "$AUTH_PROVIDER")
+    while IFS= read -r auth_file; do
+      [[ -n "$auth_file" ]] || continue
+      AUTH_FILES+=("$auth_file")
+    done < <(openclaw_live_collect_auth_files_from_csv "$AUTH_PROVIDER")
+  fi
+
+  AUTH_DIRS_CSV=""
+  if ((${#AUTH_DIRS[@]} > 0)); then
+    AUTH_DIRS_CSV="$(openclaw_live_join_csv "${AUTH_DIRS[@]}")"
+  fi
+  AUTH_FILES_CSV=""
+  if ((${#AUTH_FILES[@]} > 0)); then
+    AUTH_FILES_CSV="$(openclaw_live_join_csv "${AUTH_FILES[@]}")"
+  fi
+
+  EXTERNAL_AUTH_MOUNTS=()
+  if ((${#AUTH_DIRS[@]} > 0)); then
+    for auth_dir in "${AUTH_DIRS[@]}"; do
+      host_path="$HOME/$auth_dir"
+      if [[ -d "$host_path" ]]; then
+        EXTERNAL_AUTH_MOUNTS+=(-v "$host_path":/host-auth/"$auth_dir":ro)
+      fi
+    done
+  fi
+  if ((${#AUTH_FILES[@]} > 0)); then
+    for auth_file in "${AUTH_FILES[@]}"; do
+      host_path="$HOME/$auth_file"
+      if [[ -f "$host_path" ]]; then
+        EXTERNAL_AUTH_MOUNTS+=(-v "$host_path":/host-auth-files/"$auth_file":ro)
+      fi
+    done
+  fi
+
+  echo "==> Run ACP bind live test in Docker"
+  echo "==> Agent: $ACP_AGENT"
+  echo "==> Auth dirs: ${AUTH_DIRS_CSV:-none}"
+  echo "==> Auth files: ${AUTH_FILES_CSV:-none}"
+  docker run --rm -t \
+    -u node \
+    --entrypoint bash \
+    -e ANTHROPIC_API_KEY \
+    -e ANTHROPIC_API_KEY_OLD \
+    -e OPENCLAW_LIVE_ACP_BIND_ANTHROPIC_API_KEY="${ANTHROPIC_API_KEY:-}" \
+    -e OPENCLAW_LIVE_ACP_BIND_ANTHROPIC_API_KEY_OLD="${ANTHROPIC_API_KEY_OLD:-}" \
+    -e OPENAI_API_KEY \
+    -e COREPACK_ENABLE_DOWNLOAD_PROMPT=0 \
+    -e HOME=/home/node \
+    -e NODE_OPTIONS=--disable-warning=ExperimentalWarning \
+    -e OPENCLAW_SKIP_CHANNELS=1 \
+    -e OPENCLAW_VITEST_FS_MODULE_CACHE=0 \
+    -e OPENCLAW_DOCKER_AUTH_DIRS_RESOLVED="$AUTH_DIRS_CSV" \
+    -e OPENCLAW_DOCKER_AUTH_FILES_RESOLVED="$AUTH_FILES_CSV" \
+    -e OPENCLAW_LIVE_TEST=1 \
+    -e OPENCLAW_LIVE_ACP_BIND=1 \
+    -e OPENCLAW_LIVE_ACP_BIND_AGENT="$ACP_AGENT" \
+    -e OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND="$AGENT_COMMAND" \
+    -v "$ROOT_DIR":/src:ro \
+    -v "$CONFIG_DIR":/home/node/.openclaw \
+    -v "$WORKSPACE_DIR":/home/node/.openclaw/workspace \
+    -v "$CLI_TOOLS_DIR":/home/node/.npm-global \
+    "${EXTERNAL_AUTH_MOUNTS[@]}" \
+    "${PROFILE_MOUNT[@]}" \
+    "$LIVE_IMAGE_NAME" \
+    -lc "$LIVE_TEST_CMD"
+done
--- a/src/gateway/gateway-acp-bind.live.test.ts
+++ b/src/gateway/gateway-acp-bind.live.test.ts
@@ -65,9 +65,19 @@ function extractAssistantTexts(messages: unknown[]): string[] {
    .filter((value): value is string => typeof value === "string" && value.trim().length > 0);
 }

-function extractLastAssistantText(messages: unknown[]): string | null {
-  const texts = extractAssistantTexts(messages);
-  return texts.at(-1) ?? null;
+function createAcpRecallPrompt(liveAgent: "claude" | "codex"): string {
+  if (liveAgent === "codex") {
+    return "Please include the exact token from your immediately previous assistant reply.";
+  }
+  return "Reply with exactly the token from your immediately previous assistant reply and nothing else.";
+}
+
+function createAcpMarkerPrompt(liveAgent: "claude" | "codex", memoryNonce: string): string {
+  const token = `ACP-BIND-MEMORY-${memoryNonce}`;
+  if (liveAgent === "codex") {
+    return `Please include the exact token ${token} in your reply.`;
+  }
+  return `Reply with exactly this token and nothing else: ${token}`;
 }

 function extractSpawnedAcpSessionKey(texts: string[]): string | null {
@@ -331,6 +341,44 @@ async function sendChatAndWait(params: {
  await waitForAgentRunOk(params.client, started.runId);
 }

+async function waitForAssistantText(params: {
+  client: GatewayClient;
+  sessionKey: string;
+  contains: string;
+  minAssistantCount?: number;
+  timeoutMs?: number;
+}): Promise<{ messages: unknown[]; lastAssistantText: string }> {
+  const timeoutMs = params.timeoutMs ?? 30_000;
+  const startedAt = Date.now();
+
+  while (Date.now() - startedAt < timeoutMs) {
+    const history = await params.client.request<{ messages?: unknown[] }>("chat.history", {
+      sessionKey: params.sessionKey,
+      limit: 16,
+    });
+    const messages = history.messages ?? [];
+    const assistantTexts = extractAssistantTexts(messages);
+    const lastAssistantText = assistantTexts.at(-1) ?? null;
+    if (
+      assistantTexts.length >= (params.minAssistantCount ?? 1) &&
+      lastAssistantText?.includes(params.contains)
+    ) {
+      return { messages, lastAssistantText };
+    }
+    await sleep(500);
+  }
+
+  const finalHistory = await params.client.request<{ messages?: unknown[] }>("chat.history", {
+    sessionKey: params.sessionKey,
+    limit: 16,
+  });
+  throw new Error(
+    `timed out waiting for assistant text containing ${params.contains}: ${formatAssistantTextPreview(
+      extractAssistantTexts(finalHistory.messages ?? []),
+    )}`,
+  );
+}
+
 describeLive("gateway live (ACP bind)", () => {
  it(
    "binds a synthetic Slack DM conversation to a live ACP session and reroutes the next turn",
@@ -469,30 +517,57 @@ describeLive("gateway live (ACP bind)", () => {
        });
        logLiveStep("follow-up turn completed");

+        const firstBoundHistory = await waitForAssistantText({
+          client,
+          sessionKey: spawnedSessionKey,
+          contains: `ACP-BIND-${followupNonce}`,
+        });
+        const firstAssistantCount = extractAssistantTexts(firstBoundHistory.messages).length;
+
        await sendChatAndWait({
          client,
          sessionKey: originalSessionKey,
          idempotencyKey: `idem-memory-${randomUUID()}`,
-          message:
-            "Reply with exactly two uppercase tokens separated by a single space: " +
-            "first, the token from your immediately previous assistant reply; " +
-            `second, ACP-BIND-MEMORY-${memoryNonce}. No extra text.`,
+          message: createAcpRecallPrompt(liveAgent),
          originatingChannel: "slack",
          originatingTo: conversationId,
          originatingAccountId: accountId,
        });
-        logLiveStep("memory follow-up turn completed");
+        logLiveStep("memory recall turn completed");

-        const boundHistory = await client.request<{ messages?: unknown[] }>("chat.history", {
+        const recallHistory = await waitForAssistantText({
+          client,
          sessionKey: spawnedSessionKey,
-          limit: 16,
+          contains: `ACP-BIND-${followupNonce}`,
+          minAssistantCount: firstAssistantCount + 1,
        });
-        const assistantTexts = extractAssistantTexts(boundHistory.messages ?? []);
-        const lastAssistantText = extractLastAssistantText(boundHistory.messages ?? []);
+        const recallAssistantText = recallHistory.lastAssistantText;
+        expect(recallAssistantText).toContain(`ACP-BIND-${followupNonce}`);
+        logLiveStep("bound session transcript retained the previous token");
+        const recallAssistantCount = extractAssistantTexts(recallHistory.messages).length;
+
+        await sendChatAndWait({
+          client,
+          sessionKey: originalSessionKey,
+          idempotencyKey: `idem-marker-${randomUUID()}`,
+          message: createAcpMarkerPrompt(liveAgent, memoryNonce),
+          originatingChannel: "slack",
+          originatingTo: conversationId,
+          originatingAccountId: accountId,
+        });
+        logLiveStep("memory marker turn completed");
+
+        const boundHistory = await waitForAssistantText({
+          client,
+          sessionKey: spawnedSessionKey,
+          contains: `ACP-BIND-MEMORY-${memoryNonce}`,
+          minAssistantCount: recallAssistantCount + 1,
+        });
+        const assistantTexts = extractAssistantTexts(boundHistory.messages);
+        const lastAssistantText = boundHistory.lastAssistantText;
        expect(assistantTexts.join("\n\n")).toContain(`ACP-BIND-${followupNonce}`);
-        expect(lastAssistantText).toContain(`ACP-BIND-${followupNonce}`);
        expect(lastAssistantText).toContain(`ACP-BIND-MEMORY-${memoryNonce}`);
-        logLiveStep("bound session transcript contains follow-up token");
+        logLiveStep("bound session transcript contains the final marker token");
      } finally {
        releasePinnedPluginChannelRegistry(channelRegistry);
        clearRuntimeConfigSnapshot();