agenticdome-python-sdk 1.2.33__tar.gz → 1.2.35__tar.gz

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agenticdome-python-sdk
-Version: 1.2.33
+Version: 1.2.35
 Summary: Official Python SDK for AgenticDome security guardrails, DLP, tool authorization, and multi-agent delegation enforcement.
 Author: AgenticDome
 License-Expression: Apache-2.0
@@ -177,23 +177,41 @@ agenticdome inspect --output agenticdome-inspection.json
 
 `init` writes the secret-free inspection and local configuration. For the
 certified onboarding path, obtain the assigned sidecar and a dedicated
-`integration_copilot` key, run the tenant-bound `agenticdome plan`, attach
-and test the selected boundaries, then run
+`integration_copilot` key, run `agenticdome integrate preview`, review its
+local exact diff and change summary, explicitly approve a supported edit or
+attach other boundaries manually, then run
 `agenticdome verify --run-tests --output .agenticdome/verification.json`.
-The private plan is required because production verification consumes its
-signed, catalog-bound semantic analysis. `agenticdome scaffold` remains
-optional and writes an unapplied patch for review.
+Preview obtains the private, signed, catalog-bound semantic plan and keeps
+application source on your machine. `agenticdome integrate apply` creates a
+local Git review branch but never commits, pushes or deploys. Existing-source
+automatic edits are currently limited to one exact certified smolagents run
+pattern; all other paths need manual integration. The older `agenticdome plan`
+and `agenticdome scaffold` remain available as advanced review-only commands.
 
 For an MCP client, host, gateway or server workload, use the focused path:
 
 ```bash
 agenticdome mcp protect
-# review the detected role/transport, boundaries, registry and unapplied patch
+# Review the detected role/transport and boundaries, then preview exact local files:
+agenticdome integrate preview --target mcp
+# Review .agenticdome/scaffold/guided-integration.patch and
+# .agenticdome/scaffold/proposed/AGENTICDOME-CHANGES.md
+agenticdome integrate status
+# Optional, after approval: add review files on a local Git branch.
+agenticdome integrate apply
+# Manually attach the wrapper to the real MCP forwarder and reroute clients.
+agenticdome inspect --output agenticdome-inspection.json
 agenticdome mcp verify
 ```
 
-`mcp protect` adds no customer source changes. `mcp verify` runs the workload's
-tests, a real JSON-RPC stdio subprocess interception rehearsal and live
+`mcp protect` and `integrate preview --target mcp` do not edit customer source.
+An approved `integrate apply` adds only missing MCP review/wrapper files to a
+local Git branch; it does not rewrite an existing forwarder, route clients,
+commit or deploy. If there are no new MCP review files, apply refuses. Undo
+removes only unchanged files added by the approval. Actual interception still
+requires manually connecting the wrapper to the real request/response path,
+trusted identity and purpose, runtime credentials, client routing and tests.
+`mcp verify` runs the workload's tests, a real JSON-RPC stdio subprocess interception rehearsal and live
 assigned-sidecar decisions. The customer Control Panel combines that evidence
 with retained runtime evidence before representing production readiness.
 
@@ -216,9 +234,17 @@ agenticdome openclaw verify --output .agenticdome/verification.json
 
 `openclaw protect` fails unless the active OpenClaw runtime reports the loaded
 AgenticDome plugin, explicit consent and the exact certified typed hooks.
+The official `openclaw plugins install`, `enable`, `config set` and gateway
+restart commands are the actual runtime configuration change; the Python CLI
+does not apply or undo them. No application source patch is needed for native
+hooks. Custom skill paths outside those hooks need a reviewed manual
+`protectedExecute()` integration and workload-specific tests.
 `agenticdome plan` returns source-free, tenant-bound Integration Copilot
 guidance for review. `openclaw verify` adds workspace-test and assigned-tenant
-policy evidence. The
+policy evidence. If the same deployable workload also uses MCP, this command
+includes the MCP transport rehearsal and both `mcp_verification` and
+`openclaw_verification` in one report. Both sections must be ready before the
+report is imported; `mcp verify` alone is not a combined-workload proof. The
 customer Control Panel separately confirms tenant binding and retained
 telemetry; neither command silently modifies customer source.
 
@@ -228,6 +254,17 @@ languages and candidate boundaries can be discovered locally; business
 purpose, sensitive actions and managed-versus-sovereign deployment remain
 explicit customer decisions and are not guessed from source code.
 
+`agenticdome init` creates local configuration without contacting Copilot.
+`inspect` and `plan` send source-free structural metadata to the assigned
+tenant sidecar. A selection spanning several deployable packages is analyzed
+as bounded, sequential workload units; successful units are checkpointed
+locally so a retry does not resubmit them. The report lists analyzed parts
+and any result-list caps. It does not claim that calls or guard placement
+across separate workload units were proven; review those boundaries before
+production. If the selection exceeds the 24-unit safety limit, select the
+application or package that actually executes agent actions. No unexamined
+unit is silently omitted or reported as protected.
+
 Use **Customer Control Panel → Activate Action Firewall → Core Config** to
 choose the Developer Integration path and import the two evidence files. This
 path is for SDK, MCP, framework or custom-code integrations; Microsoft

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/README.md

@@ -64,23 +64,41 @@ agenticdome inspect --output agenticdome-inspection.json
 
 `init` writes the secret-free inspection and local configuration. For the
 certified onboarding path, obtain the assigned sidecar and a dedicated
-`integration_copilot` key, run the tenant-bound `agenticdome plan`, attach
-and test the selected boundaries, then run
+`integration_copilot` key, run `agenticdome integrate preview`, review its
+local exact diff and change summary, explicitly approve a supported edit or
+attach other boundaries manually, then run
 `agenticdome verify --run-tests --output .agenticdome/verification.json`.
-The private plan is required because production verification consumes its
-signed, catalog-bound semantic analysis. `agenticdome scaffold` remains
-optional and writes an unapplied patch for review.
+Preview obtains the private, signed, catalog-bound semantic plan and keeps
+application source on your machine. `agenticdome integrate apply` creates a
+local Git review branch but never commits, pushes or deploys. Existing-source
+automatic edits are currently limited to one exact certified smolagents run
+pattern; all other paths need manual integration. The older `agenticdome plan`
+and `agenticdome scaffold` remain available as advanced review-only commands.
 
 For an MCP client, host, gateway or server workload, use the focused path:
 
 ```bash
 agenticdome mcp protect
-# review the detected role/transport, boundaries, registry and unapplied patch
+# Review the detected role/transport and boundaries, then preview exact local files:
+agenticdome integrate preview --target mcp
+# Review .agenticdome/scaffold/guided-integration.patch and
+# .agenticdome/scaffold/proposed/AGENTICDOME-CHANGES.md
+agenticdome integrate status
+# Optional, after approval: add review files on a local Git branch.
+agenticdome integrate apply
+# Manually attach the wrapper to the real MCP forwarder and reroute clients.
+agenticdome inspect --output agenticdome-inspection.json
 agenticdome mcp verify
 ```
 
-`mcp protect` adds no customer source changes. `mcp verify` runs the workload's
-tests, a real JSON-RPC stdio subprocess interception rehearsal and live
+`mcp protect` and `integrate preview --target mcp` do not edit customer source.
+An approved `integrate apply` adds only missing MCP review/wrapper files to a
+local Git branch; it does not rewrite an existing forwarder, route clients,
+commit or deploy. If there are no new MCP review files, apply refuses. Undo
+removes only unchanged files added by the approval. Actual interception still
+requires manually connecting the wrapper to the real request/response path,
+trusted identity and purpose, runtime credentials, client routing and tests.
+`mcp verify` runs the workload's tests, a real JSON-RPC stdio subprocess interception rehearsal and live
 assigned-sidecar decisions. The customer Control Panel combines that evidence
 with retained runtime evidence before representing production readiness.
 
@@ -103,9 +121,17 @@ agenticdome openclaw verify --output .agenticdome/verification.json
 
 `openclaw protect` fails unless the active OpenClaw runtime reports the loaded
 AgenticDome plugin, explicit consent and the exact certified typed hooks.
+The official `openclaw plugins install`, `enable`, `config set` and gateway
+restart commands are the actual runtime configuration change; the Python CLI
+does not apply or undo them. No application source patch is needed for native
+hooks. Custom skill paths outside those hooks need a reviewed manual
+`protectedExecute()` integration and workload-specific tests.
 `agenticdome plan` returns source-free, tenant-bound Integration Copilot
 guidance for review. `openclaw verify` adds workspace-test and assigned-tenant
-policy evidence. The
+policy evidence. If the same deployable workload also uses MCP, this command
+includes the MCP transport rehearsal and both `mcp_verification` and
+`openclaw_verification` in one report. Both sections must be ready before the
+report is imported; `mcp verify` alone is not a combined-workload proof. The
 customer Control Panel separately confirms tenant binding and retained
 telemetry; neither command silently modifies customer source.
 
@@ -115,6 +141,17 @@ languages and candidate boundaries can be discovered locally; business
 purpose, sensitive actions and managed-versus-sovereign deployment remain
 explicit customer decisions and are not guessed from source code.
 
+`agenticdome init` creates local configuration without contacting Copilot.
+`inspect` and `plan` send source-free structural metadata to the assigned
+tenant sidecar. A selection spanning several deployable packages is analyzed
+as bounded, sequential workload units; successful units are checkpointed
+locally so a retry does not resubmit them. The report lists analyzed parts
+and any result-list caps. It does not claim that calls or guard placement
+across separate workload units were proven; review those boundaries before
+production. If the selection exceeds the 24-unit safety limit, select the
+application or package that actually executes agent actions. No unexamined
+unit is silently omitted or reported as protected.
+
 Use **Customer Control Panel → Activate Action Firewall → Core Config** to
 choose the Developer Integration path and import the two evidence files. This
 path is for SDK, MCP, framework or custom-code integrations; Microsoft

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/agenticdome_python_sdk.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agenticdome-python-sdk
-Version: 1.2.33
+Version: 1.2.35
 Summary: Official Python SDK for AgenticDome security guardrails, DLP, tool authorization, and multi-agent delegation enforcement.
 Author: AgenticDome
 License-Expression: Apache-2.0
@@ -177,23 +177,41 @@ agenticdome inspect --output agenticdome-inspection.json
 
 `init` writes the secret-free inspection and local configuration. For the
 certified onboarding path, obtain the assigned sidecar and a dedicated
-`integration_copilot` key, run the tenant-bound `agenticdome plan`, attach
-and test the selected boundaries, then run
+`integration_copilot` key, run `agenticdome integrate preview`, review its
+local exact diff and change summary, explicitly approve a supported edit or
+attach other boundaries manually, then run
 `agenticdome verify --run-tests --output .agenticdome/verification.json`.
-The private plan is required because production verification consumes its
-signed, catalog-bound semantic analysis. `agenticdome scaffold` remains
-optional and writes an unapplied patch for review.
+Preview obtains the private, signed, catalog-bound semantic plan and keeps
+application source on your machine. `agenticdome integrate apply` creates a
+local Git review branch but never commits, pushes or deploys. Existing-source
+automatic edits are currently limited to one exact certified smolagents run
+pattern; all other paths need manual integration. The older `agenticdome plan`
+and `agenticdome scaffold` remain available as advanced review-only commands.
 
 For an MCP client, host, gateway or server workload, use the focused path:
 
 ```bash
 agenticdome mcp protect
-# review the detected role/transport, boundaries, registry and unapplied patch
+# Review the detected role/transport and boundaries, then preview exact local files:
+agenticdome integrate preview --target mcp
+# Review .agenticdome/scaffold/guided-integration.patch and
+# .agenticdome/scaffold/proposed/AGENTICDOME-CHANGES.md
+agenticdome integrate status
+# Optional, after approval: add review files on a local Git branch.
+agenticdome integrate apply
+# Manually attach the wrapper to the real MCP forwarder and reroute clients.
+agenticdome inspect --output agenticdome-inspection.json
 agenticdome mcp verify
 ```
 
-`mcp protect` adds no customer source changes. `mcp verify` runs the workload's
-tests, a real JSON-RPC stdio subprocess interception rehearsal and live
+`mcp protect` and `integrate preview --target mcp` do not edit customer source.
+An approved `integrate apply` adds only missing MCP review/wrapper files to a
+local Git branch; it does not rewrite an existing forwarder, route clients,
+commit or deploy. If there are no new MCP review files, apply refuses. Undo
+removes only unchanged files added by the approval. Actual interception still
+requires manually connecting the wrapper to the real request/response path,
+trusted identity and purpose, runtime credentials, client routing and tests.
+`mcp verify` runs the workload's tests, a real JSON-RPC stdio subprocess interception rehearsal and live
 assigned-sidecar decisions. The customer Control Panel combines that evidence
 with retained runtime evidence before representing production readiness.
 
@@ -216,9 +234,17 @@ agenticdome openclaw verify --output .agenticdome/verification.json
 
 `openclaw protect` fails unless the active OpenClaw runtime reports the loaded
 AgenticDome plugin, explicit consent and the exact certified typed hooks.
+The official `openclaw plugins install`, `enable`, `config set` and gateway
+restart commands are the actual runtime configuration change; the Python CLI
+does not apply or undo them. No application source patch is needed for native
+hooks. Custom skill paths outside those hooks need a reviewed manual
+`protectedExecute()` integration and workload-specific tests.
 `agenticdome plan` returns source-free, tenant-bound Integration Copilot
 guidance for review. `openclaw verify` adds workspace-test and assigned-tenant
-policy evidence. The
+policy evidence. If the same deployable workload also uses MCP, this command
+includes the MCP transport rehearsal and both `mcp_verification` and
+`openclaw_verification` in one report. Both sections must be ready before the
+report is imported; `mcp verify` alone is not a combined-workload proof. The
 customer Control Panel separately confirms tenant binding and retained
 telemetry; neither command silently modifies customer source.
 
@@ -228,6 +254,17 @@ languages and candidate boundaries can be discovered locally; business
 purpose, sensitive actions and managed-versus-sovereign deployment remain
 explicit customer decisions and are not guessed from source code.
 
+`agenticdome init` creates local configuration without contacting Copilot.
+`inspect` and `plan` send source-free structural metadata to the assigned
+tenant sidecar. A selection spanning several deployable packages is analyzed
+as bounded, sequential workload units; successful units are checkpointed
+locally so a retry does not resubmit them. The report lists analyzed parts
+and any result-list caps. It does not claim that calls or guard placement
+across separate workload units were proven; review those boundaries before
+production. If the selection exceeds the 24-unit safety limit, select the
+application or package that actually executes agent actions. No unexamined
+unit is silently omitted or reported as protected.
+
 Use **Customer Control Panel → Activate Action Firewall → Core Config** to
 choose the Developer Integration path and import the two evidence files. This
 path is for SDK, MCP, framework or custom-code integrations; Microsoft

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/agenticdome_python_sdk.egg-info/SOURCES.txt

@@ -28,6 +28,7 @@ agenticdome_sdk/copilot_ir_collector.cjs
 agenticdome_sdk/crewai.py
 agenticdome_sdk/demo.py
 agenticdome_sdk/google_adk.py
+agenticdome_sdk/guided_integration.py
 agenticdome_sdk/hook_catalog.py
 agenticdome_sdk/identity.py
 agenticdome_sdk/langgraph.py

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/agenticdome_sdk/catalog_snapshot.json

@@ -2,12 +2,12 @@
     "schema": "agenticdome.catalog-snapshot.v1",
     "catalog_schema": "agenticdome.hook-catalog.v1",
     "verified_at": "2026-09-26",
-    "python_sdk_version": "1.2.33",
+    "python_sdk_version": "1.2.35",
     "published_packages": {
         "agenticdome-python-sdk": {
             "registry": "pypi",
-            "version": "1.2.33",
-            "url": "https://pypi.org/project/agenticdome-python-sdk/1.2.33/"
+            "version": "1.2.35",
+            "url": "https://pypi.org/project/agenticdome-python-sdk/1.2.35/"
         },
         "agenticdome-sdk": {
             "registry": "npm",

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/agenticdome_sdk/connect.py

@@ -17,9 +17,6 @@ from pathlib import Path
 from typing import Any, Dict, Optional
 
 from .attestation import ensure_attestation_key
-from .hook_catalog import CATALOG_SCHEMA, CATALOG_VERIFIED_AT, catalog_digest
-
-
 SCOPES = ["onboarding:read", "metadata:write", "evidence:write", "workload:write", "repository:prepare"]
 
 
@@ -124,7 +121,10 @@ def _approve_metadata(ir: Dict[str, Any], assume_yes: bool) -> None:
 
 def run_connect(root: Path, args: Any) -> Dict[str, Any]:
     # Import lazily to avoid a module cycle and keep the public client purely orchestration code.
-    from .onboarding_cli import _installed_sdk_version, inspect_repository
+    from .onboarding_cli import (
+        _active_copilot_catalog_binding, _copilot_semantic_analysis,
+        _ensure_workload_id, _installed_sdk_version, _scope_gap_message, inspect_repository,
+    )
 
     portal = str(args.portal).rstrip("/")
     if not portal.startswith("https://") and not getattr(args, "allow_insecure_http", False):
@@ -155,12 +155,23 @@ def run_connect(root: Path, args: Any) -> Dict[str, Any]:
     if not session.get("runtime") or not session.get("copilot"):
         raise SystemExit("Authentication succeeded, but no eligible runtime is assigned. The saved portal session can be resumed after an administrator enables an assignment policy.")
 
-    report = inspect_repository(root)
+    local_workload_id = _ensure_workload_id(root)
+    report = inspect_repository(root, remote_analysis=False)
+    if report.get("scope", {}).get("complete") is not True:
+        raise SystemExit(_scope_gap_message(report))
     ir = report["copilot_ir"]
     _approve_metadata(ir, bool(args.yes))
     private_pem, public_pem = ensure_attestation_key(root / ".agenticdome")
     del private_pem
-    workload_uuid = str(uuid.uuid5(uuid.NAMESPACE_URL, session["tenant_id"] + ":" + root.name))
+    workload_uuid = str(uuid.uuid5(uuid.NAMESPACE_URL, session["tenant_id"] + ":" + local_workload_id))
+    previous_session_path = root / ".agenticdome" / "connect-session.json"
+    if previous_session_path.is_file():
+        try:
+            previous_session = json.loads(previous_session_path.read_text(encoding="utf-8"))
+            if previous_session.get("tenant_id") == session["tenant_id"]:
+                workload_uuid = str(uuid.UUID(str(previous_session["workload_uuid"])))
+        except (OSError, ValueError, KeyError, TypeError, AttributeError, json.JSONDecodeError):
+            raise SystemExit("The previous Connect workload identity is unreadable. Restore its connect-session.json before retrying to avoid creating a duplicate workload.")
     frameworks = [item["key"] for item in report.get("frameworks", [])]
     _request(portal + "/api/agentguard/connect/workloads", {
         "session_uuid": session["session_uuid"], "workload_uuid": workload_uuid,
@@ -171,38 +182,51 @@ def run_connect(root: Path, args: Any) -> Dict[str, Any]:
 
     runtime_base = str(session["runtime"]["api_base"]).rstrip("/")
     copilot_key = str(session["copilot"]["api_key"])
-    ir_digest = hashlib.sha256(json.dumps(ir, sort_keys=True, separators=(",", ":")).encode("utf-8")).hexdigest()
-    plan = _request(runtime_base + "/integration-copilot/v1/analyze", {
-        "schema": "agenticdome.copilot-request.v1", "ir": ir,
-        "catalog_binding": {"schema": CATALOG_SCHEMA, "catalog_digest": catalog_digest(), "verified_at": CATALOG_VERIFIED_AT},
-    }, headers={"X-API-Key": copilot_key, "X-Tenant-Id": session["tenant_id"], "Idempotency-Key": hashlib.sha256((session["tenant_id"] + ir_digest + catalog_digest()).encode()).hexdigest()})
-    if plan.get("schema") != "agenticdome.copilot-plan.v1" or plan.get("tenant_id") != session["tenant_id"]:
-        raise SystemExit("The assigned private Copilot returned an invalid tenant-bound plan.")
-    semantic = plan.get("semantic_analysis") or {}
-    if semantic.get("ir_sha256") != ir_digest:
-        raise SystemExit("The private Copilot plan was not bound to the submitted metadata.")
-    _request(portal + "/api/agentguard/connect/sessions/" + session["session_uuid"] + "/metadata", {"ir": ir, "signed_plan": plan}, session["access_token"])
-
+    from .onboarding_cli import _ir_sha256
+    ir_digest = _ir_sha256(ir)
     state_dir = root / ".agenticdome"
     state_dir.mkdir(parents=True, exist_ok=True)
-    cache = {"schema": "agenticdome.copilot-cache.v1", "source_upload": False, "tenant_id": session["tenant_id"], "api_base": runtime_base, "semantic_analysis": semantic, "catalog_binding": plan.get("catalog_binding")}
-    (state_dir / "copilot-analysis.json").write_text(json.dumps(cache, indent=2, sort_keys=True) + "\n", encoding="utf-8")
     config_path = state_dir / "config.json"
     if not config_path.exists():
         config_path.write_text(json.dumps({
-            "schema": "agenticdome.project-config.v1", "frameworks": frameworks or ["custom-python"],
+            "schema": "agenticdome.project-config.v1", "workload_id": local_workload_id,
+            "frameworks": frameworks or ["custom-python"],
             "execution_broker_mode": "policy",
             "business_purpose": "REVIEW_REQUIRED_NOT_INVENTED", "sensitive_tools": [],
             "deployment": {"preference": "managed", "region": session["runtime"].get("region") or "auto"},
             "source_upload": False,
         }, indent=2, sort_keys=True) + "\n", encoding="utf-8")
-    resumable = {"schema": "agenticdome.connect-state.v1", "portal": portal, "tenant_id": session["tenant_id"], "session_uuid": session["session_uuid"], "workload_uuid": workload_uuid, "token_expires_at": session["expires_at"], "access_token": session["access_token"]}
-    state_path = state_dir / "connect-session.json"
-    state_path.write_text(json.dumps(resumable, indent=2, sort_keys=True) + "\n", encoding="utf-8")
-    os.chmod(state_path, 0o600)
     previous = {name: os.environ.get(name) for name in ("AGENTICDOME_API_BASE", "AGENTICDOME_COPILOT_API_KEY", "AGENTICDOME_TENANT_ID")}
     os.environ.update({"AGENTICDOME_API_BASE": runtime_base, "AGENTICDOME_COPILOT_API_KEY": copilot_key, "AGENTICDOME_TENANT_ID": session["tenant_id"]})
     try:
+        semantic = _copilot_semantic_analysis(root, ir, required=True)
+        binding = _active_copilot_catalog_binding(root)
+        if semantic.get("ir_sha256") != ir_digest or not binding:
+            raise SystemExit("The private Copilot result was not bound to this workload and the installed SDK catalog.")
+        # The portal stores a compact inventory, never the potentially large
+        # per-function graph. Every selected part was analyzed at the sidecar.
+        portal_ir = {
+            "schema": "agenticdome.copilot-ir.v1", "source_upload": False,
+            "metadata_kind": "bounded_workload_summary", "selected_ir_sha256": ir_digest,
+            "files": [], "functions": [],
+            "features": {"functions": len(ir.get("functions", []))},
+            "scope": {
+                "complete": True, "fingerprint": ir.get("scope", {}).get("fingerprint"),
+                "eligible_files_scanned": ir.get("scope", {}).get("eligible_files_scanned", 0),
+            },
+            "workload_coverage": {
+                "selected_parts": semantic.get("workload_coverage", {}).get("selected_parts", 1),
+                "analyzed_parts": semantic.get("workload_coverage", {}).get("analyzed_parts", 1),
+                "cross_part_flow_proven": semantic.get("workload_coverage", {}).get("cross_part_flow_proven", True),
+            },
+        }
+        plan = {
+            "schema": "agenticdome.copilot-plan.v1", "tenant_id": session["tenant_id"],
+            "semantic_analysis": semantic, "catalog_binding": binding,
+            "aggregation": {"kind": "sdk_local_bounded_parts", "signature_scope": "individual_sidecar_parts_only"},
+        }
+        _request(portal + "/api/agentguard/connect/sessions/" + session["session_uuid"] + "/metadata", {"ir": portal_ir, "signed_plan": plan}, session["access_token"])
+
         from .onboarding_cli import create_scaffold
         patch_path = create_scaffold(root)
     finally:
@@ -211,6 +235,11 @@ def run_connect(root: Path, args: Any) -> Dict[str, Any]:
                 os.environ.pop(name, None)
             else:
                 os.environ[name] = value
+
+    resumable = {"schema": "agenticdome.connect-state.v1", "portal": portal, "tenant_id": session["tenant_id"], "session_uuid": session["session_uuid"], "workload_uuid": workload_uuid, "token_expires_at": session["expires_at"], "access_token": session["access_token"]}
+    state_path = state_dir / "connect-session.json"
+    state_path.write_text(json.dumps(resumable, indent=2, sort_keys=True) + "\n", encoding="utf-8")
+    os.chmod(state_path, 0o600)
     pull_request = None
     if args.open_pr:
         if not args.repository_connection:
@@ -222,6 +251,15 @@ def run_connect(root: Path, args: Any) -> Dict[str, Any]:
         "next_action": "Ask AgenticDome to confirm policy-aware sidecar readiness before live verification.",
     })
     broker_setup["sdk_mode"] = effective_config.get("execution_broker_mode", "existing_sdk_or_environment_default")
+    cross_part_review_required = semantic.get("workload_coverage", {}).get("cross_part_flow_proven") is False
+    next_action = (
+        "Review the generated pull request and run agenticdome verify. Then review cross-part call paths and workload-specific tests in the portal before activation."
+        if args.open_pr else
+        "Review the generated patch, apply it on a branch, and run agenticdome verify. Then review cross-part call paths and workload-specific tests in the portal before activation."
+    ) if cross_part_review_required else (
+        "Review the generated pull request and run agenticdome verify."
+        if args.open_pr else "Review the generated patch, apply it on a branch, then run agenticdome verify."
+    )
     return {
         "status": "changeset_ready", "source_upload": False, "tenant_id": session["tenant_id"],
         "session_uuid": session["session_uuid"], "workload_uuid": workload_uuid,
@@ -229,9 +267,10 @@ def run_connect(root: Path, args: Any) -> Dict[str, Any]:
         "execution_broker": broker_setup,
         "attachment_points": len(semantic.get("attachment_points") or []),
         "bypass_risks": len(semantic.get("bypass_risks") or []),
+        "cross_part_review_required": cross_part_review_required,
         "patch": str(patch_path.relative_to(root)),
         "pull_request": pull_request,
-        "next_action": "Review the generated pull request and run agenticdome verify." if pull_request else "Review the generated patch, apply it on a branch, then run agenticdome verify.",
+        "next_action": next_action,
         "sdk_version": _installed_sdk_version(),
     }
 

{agenticdome_python_sdk-1.2.33 → agenticdome_python_sdk-1.2.35}/agenticdome_sdk/copilot_ir.py

@@ -350,11 +350,13 @@ def _collect_typescript(root: Path, paths: Sequence[Path]) -> Tuple[List[Dict[st
 def _typescript_fallback(root: Path, paths: Sequence[Path], reason: str) -> Tuple[List[Dict[str, Any]], Dict[str, Any]]:
     functions: List[Dict[str, Any]] = []
     parsed = 0
+    read_errors = 0
     call_pattern = re.compile(r"\b([A-Za-z_$][\w$]*(?:\s*\.\s*[A-Za-z_$][\w$]*)*)\s*\(")
     for path in paths:
         try:
             lines = path.read_text(encoding="utf-8", errors="ignore").splitlines()
         except OSError:
+            read_errors += 1
             continue
         events = []
         for line_number, line in enumerate(lines, start=1):
@@ -368,7 +370,7 @@ def _typescript_fallback(root: Path, paths: Sequence[Path], reason: str) -> Tupl
         "engine": "typescript-structural-fallback",
         "available": False,
         "files_parsed": parsed,
-        "parse_errors": 0,
+        "parse_errors": read_errors,
         "reason": reason,
         "claim": "Generic structural collection only; install TypeScript for compiler AST evidence",
     }
@@ -382,7 +384,10 @@ def collect_repository_ir(root: Path, paths: Sequence[Path]) -> Dict[str, Any]:
     python_functions, python_engine = _collect_python(root, python_paths)
     typescript_functions, typescript_engine = _collect_typescript(root, typescript_paths)
     functions = python_functions + typescript_functions
-    complete = len(paths) <= MAX_IR_FILES and len(functions) <= MAX_IR_FUNCTIONS
+    collection_errors = int(python_engine.get("parse_errors", 0)) + int(typescript_engine.get("parse_errors", 0))
+    files_parsed = int(python_engine.get("files_parsed", 0)) + int(typescript_engine.get("files_parsed", 0))
+    complete = (len(paths) <= MAX_IR_FILES and len(functions) <= MAX_IR_FUNCTIONS
+                and collection_errors == 0 and files_parsed == len(paths))
     return {
         "schema": IR_SCHEMA,
         "source_upload": False,
@@ -394,7 +399,8 @@ def collect_repository_ir(root: Path, paths: Sequence[Path]) -> Dict[str, Any]:
             "files_selected": len(selected),
             "symbols_found": len(functions),
             "complete": complete,
-            "limit_reason": "file_limit" if len(paths) > MAX_IR_FILES else "symbol_limit" if len(functions) > MAX_IR_FUNCTIONS else None,
+            "limit_reason": "file_limit" if len(paths) > MAX_IR_FILES else "symbol_limit" if len(functions) > MAX_IR_FUNCTIONS
+                else "parse_or_read_error" if collection_errors or files_parsed != len(paths) else None,
         },
         "privacy": {
             "source_text": False,