mirror of
https://github.com/rennf93/roboco.git
synced 2026-08-03 07:23:24 +02:00
* feat(a2a): deliver latest incoming message preview into the claim briefing
list_unread_a2a now carries last_message_preview (the latest message from the
OTHER agent, never the agent's own reply), fetched via a correlated subquery in
the same query — no N+1 on the per-verb briefing path.
* feat(a2a): read_a2a verb delivers unread message bodies to the agent
A2AService.get_unread_messages returns the caller's unread INCOMING messages
(never its own sends), marking exactly those rows read atomically so a message
arriving mid-call is preserved. Wired as the read_a2a content verb (route +
do_server tool + granted to every delivery role) — the content-bearing read the
A2A inbox lacked (read_messages only zeroed the counter).
* docs(rag): document read_a2a as the A2A content-read path
* fix(task): backlog activation no longer requires a discussion session
Removes the SessionTaskTable gate in activate() (and its dangling log field),
deletes _inherit_parent_session + its create() call, and drops the now-unused
SessionTaskTable import. Coordination rides task state; the session subsystem is
being retired. Tests updated to the new (no-session) behavior.
* fix(orchestrator): drop session sweep from _run_sweep
Removes the messaging import + sweep_timed_out_sessions call. That import sat
outside the try/except, so once messaging.py is deleted it would have killed the
entire sweep cascade (budget kill-switch, token rollups, retention, image prune,
superseded-PR reconcile). Notification sweep + all maintenance sweeps unchanged.
* release-manager --no-tags read-clone fix
* test: update evidence_repo unit test for a2a last_message_preview
* refactor(gateway): drop session propagation on delegate
Removes propagate_sessions_to_subtask from delegate(), the ChoreographerDeps
messaging field + property, and the ChoreographerDeps messaging arg in deps.py
(ContentActions messaging + import stay until the verbs are removed). Deletes the
propagation test; strips the now-invalid messaging kwarg from ChoreographerDeps
test builders.
* refactor(gateway): remove say/open_session/link_session/channels verbs
Removes the four channel/session verbs across content_actions (impls +
ContentActionsDeps.messaging), do_server (tools + registry), role_config (grants
+ _CHANNEL_DISCOVERY), do.py (routes), schemas/v1/do.py (request models), and
deps.py (MessagingService import + construction). Regenerates the prompt verb
tables. dm/notify/read_messages/read_a2a stay. Tests deleted/updated accordingly.
* uv.lock Upgrade
* refactor: remove conversation RAG indexing; Secretary announces via notification
Drops the CONVERSATIONS index (index_conversation, ConversationsIndexPlugin,
IndexType.CONVERSATIONS enum, IndexConversationParams, mentor.py type-label, the
messaging index hook) and its chunk-table manifest entries. The Secretary's
ANNOUNCE/RELAY_MESSAGE now fan out a BROADCAST notification to every agent's
inbox (NotificationService.broadcast) instead of posting to a dead channel.
* fix(panel): label RAG health error lines by subsystem
A red llm_error (e.g. the glm-5.2:cloud weekly-limit 429) rendered under
the 'Embedding: ok' header with no label, reading as an embedding failure.
Prefix each error line with LLM / Embedding / Vector store.
* refactor: remove channel/message reads from metrics, dashboard, git, events
MetricsService drops get_communication_volume + the MessageTable
message-count in get_agent_metrics (and the now-dead messages_sent_week
field). DashboardService drops get_channel_feeds/_compute_channel_status
and the message read in get_recent_activity (task activity kept);
get_auditor_metrics no longer reports communication_volume.
GitService's two primary-session-id helpers always return None now
(callers already treat None as "no primary session"). events/handlers.py
drops the SESSION_CLOSED/SESSION_TIMEOUT subscriptions + the
handle_session_boundary handler.
Forced follow-on: api/routes/dashboard.py + api/schemas/dashboard.py
dropped the now-dangling live_feeds/ChannelFeed surface and the
/metrics/communication route, which wrapped the removed service calls
directly (mypy would otherwise fail on the missing attributes).
* refactor: delete MessagingService + channel seeding
Edited db/__init__.py and services/__init__.py first (drop the unconditional
Channel/Group/Message/Session table + MessagingService re-exports), then
deleted services/messaging.py, then trimmed db/seed.py to only create_agents
(create_channels/create_channel_memberships/create_initial_messages gone).
Forced expansion: api/routes/{channels,groups,sessions,messages}.py import
roboco.services.messaging directly (not through the package __init__), as
does api/routes/tasks.py (the session-links embed on GET /tasks/{id} and the
GET /{id}/sessions route). Deleting messaging.py without addressing these
breaks `import roboco.api.app` immediately, since app.py eagerly imports all
route modules at startup. Since the 4 CRUD route files are 100%
MessagingService-backed with zero independent logic (and are wholesale
deletes in the plan's later API-routes task anyway), deleted them now +
unmounted from app.py/routes/__init__.py; tasks.py got the same surgical
trim its later task already specified (drop session-links embed +
TaskSessionLinkResponse/TaskResponse.sessions). This pulls a slice of that
later work forward — the routes/schemas for channels/groups/sessions/messages
still need their own pass, but their messaging-coupled parts are gone.
Verified with a full-suite collection sweep (12010 tests collected, zero
import errors) beyond the directly touched test dirs, given the expanded
blast radius.
* refactor: remove channel/session/message models, tables, and channel policy
Models: deleted channel.py/group.py/session.py/messaging.py wholesale
(zero external consumers besides the models/__init__.py re-export).
message.py surgically trimmed: removed MessageCreate (dead) and MessageEdit
(never instantiated; ExtractedMessage.edit_history retyped to
list[dict[str, Any]] to match how it's actually persisted — confirmed
ExtractedMessage was never written to any DB table, so MessageTable's
removal carries no functional risk to the kept extraction pipeline).
base.py: removed SessionStatus + ChannelType, kept MessageType. Also
removed the confirmed-dead channels_read/channels_write fields from
models/agent.py:AgentPermissions and models/dashboard.py:ChannelFeedData.
db/tables.py: deleted ChannelTable/GroupTable/SessionTable/SessionTaskTable/
MessageTable, TaskTable.session_links, and JournalEntryTable.session_id —
cascaded through models/journal.py, services/journal.py, and
api/schemas+routes/journals.py (22 plumbing sites).
foundation/policy/communications.py: removed the ChannelSpec/CHANNELS
catalog + TEAM_SCOPED_ROLES/_CELL_*/_AUDITOR_ONLY helpers, kept the
notification policy (Priority/parse_priority/NOTIFY_SENDER_ROLES/
ACK_REQUIRED_BY_TYPE). enforcement/channel_access.py deleted (confirmed
fully dead in production). agents_config.py: removed CHANNEL_ACCESS
(kept A2A_ALLOWED_PAIRS). seeds/initial_data.py: removed
DEFAULT_CHANNELS/CHANNEL_MEMBERSHIPS/AUDITOR_SILENT_ACCESS + the
never-consumed INITIAL_MESSAGES. config.py: removed
session_idle_timeout_seconds (zero consumers). exceptions.py: removed
dead ChannelError/ChannelAccessDeniedError/SessionClosedError.
Forced expansion beyond the original file list — ChannelType cascaded
into a live, mounted surface the plan didn't trace: agents_config.
CHANNEL_ACCESS -> services/permissions.py's channel-RBAC methods (not
models/permissions.py, which turned out to have no channel code at all)
-> two real endpoints in api/routes/stream.py (GET /permissions,
GET /permissions/channel/{name}) and two dependency factories in
api/deps.py. Removed the channel methods + fields, deleted the
channel-specific stream.py endpoint, deleted require_channel_read/write.
Also deleted api/schemas/{channels,sessions}.py (hard dependency on the
removed enums; already fully dead after the Task 10 route deletions) and
api/schemas/messages.py (a TYPE_CHECKING-only import of the deleted
MessageTable; likewise already fully dead) + its dedicated test file.
Test updates: test_permissions.py -14 channel tests (matches the planned
count exactly), test_communications.py / test_communications_consumers.py
split to keep only notification-policy coverage, test_exceptions.py -9,
test_deps.py -4, plus the journal/stream/foundation-smoke fallout. Also
fixed a pre-existing (Task 7) broken assertion in
test_foundation_phase3_smoke.py that inspected a `say()` method already
removed from ContentActions.
Verified: full-suite collection (11961 tests, zero import errors) and a
complete test run (11567 passed, 394 skipped, 0 failed) in addition to
the targeted suites.
* migration: drop channels/groups/sessions/session_tasks/messages + enum types
alembic/versions/060_drop_messaging.py: drop_column journal_entries.
session_id (sidesteps hardcoding the FK constraint name — verified
empirically against a live migrated DB that it's actually
fk_journal_entries_session_id_sessions, but drop_column doesn't care
either way); drop_table in FK order (messages -> session_tasks ->
sessions -> groups -> channels); DROP TABLE IF EXISTS chunks_conversations
(runtime-provisioned, not alembic-managed, would otherwise orphan); DROP
TYPE IF EXISTS for messagetype/sessionstatus/sessionscope/channeltype
(messagetype's Python enum stays for ExtractedMessage, but the DB type
had zero live columns left once MessageTable was dropped in the prior
commit). downgrade() raises NotImplementedError — one-way removal.
Pruned scripts/reset_runtime_state.sql + .sh: removed the DELETE/COUNT
lines for messages/session_tasks/sessions/groups/channels and the
groups.active_session_id reset block.
Verified end-to-end against a scratch Postgres DB: full migration chain
001->060 applies cleanly, alembic heads shows a single head, all 6 dropped
tables + 4 enum types + the journal_entries.session_id column are
confirmed gone, journal_entries keeps only its journal_id/task_id FKs,
downgrade correctly raises NotImplementedError without corrupting DB
state, and the pruned reset_runtime_state.sql runs clean (no errors)
against a fully-migrated DB.
* refactor(api): remove channel/session/message routes + WS streams
Most of this task's file list was already forced through in earlier
commits (routes/{channels,groups,sessions,messages}.py + app.py/__init__.py
unmounting in the MessagingService-deletion commit; tasks.py's
session-links embed + GET /{id}/sessions + schemas/tasks.py's
TaskResponse.sessions in that same commit; deps.py's require_channel_read/
write + schemas/{channels,sessions}.py in the models/tables commit). This
closes out what was left:
- api/websocket.py: deleted the channel_stream + session_stream routes,
ConnectionManager's channel_connections/session_connections dicts,
connect_channel/connect_session, broadcast_to_channel/broadcast_to_session,
get_channel_subscriber_count, and their cleanup lines in disconnect().
Agent streams, notification streams, and the operator system stream are
untouched.
- api/websocket_bridge.py: deleted _handle_session_event +
_handle_message_event and their SESSION_CREATED/SESSION_CLOSED/
SESSION_TIMEOUT/MESSAGE_SENT subscriptions. The A2A live-view, rate-limit,
usage, agent-lifecycle, and notification bridges are untouched.
- api/schemas/websocket.py: removed NewMessageBroadcast, WSMessageNew,
WSMessageEdit, WSMessageDelete, WSSessionClosed — kept the WSMessage base
class (still subclassed by the kept WSAgentStream/WSNotification) plus
those two.
- api/schemas/groups.py: deleted (already fully orphaned since routes/
groups.py was removed; its GroupResponse/GroupDetailResponse had zero
consumers).
Updated the 5 websocket test files accordingly (removed the channel/
session-specific tests + fixed imports); test_websocket_bridge.py's
registration-coverage test dropped the SESSION_*/MESSAGE_SENT assertions.
Verified: full-suite collection (11943 tests, zero import errors) and a
complete test run (11549 passed, 394 skipped, 0 failed).
* docs: retire channels/sessions/messages from agent-facing docs + CLAUDE.md
Rewrites docs/rag (RAG-indexed) + docs/map + CLAUDE.md to reflect A2A (dm +
read_a2a) as primary agent comms; deletes the channel docs, splits messaging-tools
+ messaging-notification (renamed notification.md), swaps the WS worked example to
A2A_MESSAGE_SENT. _complete_map.md still needs regeneration (generated file).
* refactor(panel): remove Communications surface (channels/sessions)
Deletes the /communications routes, message components, task-detail Sessions tab,
use-channels + channel/session WS hooks, and the channels/sessions/messages/groups
api clients; prunes the Channel/Session/Message/Group types + mock data. (Auditor
live-feeds + dashboard.ts dead-route cleanup is a follow-up.)
* refactor(panel): drop auditor channel-feed + dead communication-metric route
* docs(map): regenerate _complete_map from updated slices
* fix(a2a): reduce get_unread_messages complexity below xenon C + stale comments
Extract the per-conversation unread-counter recompute into _reset_unread_counter
(the CI quality gate flagged get_unread_messages as rank C). Also drop the deleted
open_session from a content_actions comment and reword an evidence_repo docstring
that cited the removed messaging._notify_mentions.
---------
Co-authored-by: Renn F <rennf93@users.noreply.github.com>
802 lines
29 KiB
Python
802 lines
29 KiB
Python
"""flow_server wires gateway rejections into the SDK per-verb circuit breaker.
|
|
|
|
Phase 3 Task 14 added the SDK-side tracker (POST /verb/attempted,
|
|
GET /verb/circuit_status, Envelope.circuit_open). This module verifies
|
|
the *missing wiring*: rejection envelopes from the gateway must be
|
|
forwarded to the SDK, and when the breaker opens the rejection must be
|
|
substituted with the circuit_open envelope before reaching the agent.
|
|
|
|
The tests stub the orchestrator's httpx.Client (path '/api/v1/flow/...')
|
|
and the SDK's httpx.Client (path '/verb/attempted') so the helper can
|
|
be exercised end-to-end without a real network. We pick which mock to
|
|
return by inspecting the URL the code under test is hitting.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import importlib
|
|
import json
|
|
from typing import TYPE_CHECKING, Any
|
|
from unittest.mock import MagicMock, patch
|
|
|
|
import pytest
|
|
|
|
if TYPE_CHECKING:
|
|
import types
|
|
from pathlib import Path
|
|
|
|
|
|
_FULL_MANIFEST = {
|
|
"agent_id": "00000000-0000-0000-0000-000000000099",
|
|
"role": "developer",
|
|
"team": "backend",
|
|
"workspace_path": "/tmp/test",
|
|
"flow_tools": [
|
|
"give_me_work",
|
|
"i_will_work_on",
|
|
"open_pr",
|
|
"i_am_done",
|
|
"i_am_blocked",
|
|
"unclaim",
|
|
"resume",
|
|
"i_am_idle",
|
|
"claim_review",
|
|
"pass",
|
|
"fail",
|
|
"claim_doc_task",
|
|
"i_documented",
|
|
"triage",
|
|
"triage_all",
|
|
"unblock",
|
|
"complete",
|
|
"escalate_up",
|
|
"i_will_plan",
|
|
"delegate",
|
|
"submit_up",
|
|
"escalate_to_ceo",
|
|
],
|
|
"do_tools": ["commit", "note", "dm", "evidence"],
|
|
"read_tools": ["Read", "Glob", "Grep"],
|
|
"write_tools": ["Edit", "Write"],
|
|
"bash_allowed": True,
|
|
"subagent_allowed": False,
|
|
"subagent_model": None,
|
|
"env": {},
|
|
}
|
|
|
|
|
|
@pytest.fixture()
|
|
def flow_module(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> types.ModuleType:
|
|
"""Import flow_server with a tmp manifest + known orchestrator/SDK URLs."""
|
|
manifest_path = tmp_path / "tool-manifest.json"
|
|
manifest_path.write_text(json.dumps(_FULL_MANIFEST))
|
|
|
|
monkeypatch.setenv("ROBOCO_AGENT_ID", "00000000-0000-0000-0000-000000000099")
|
|
monkeypatch.setenv("ROBOCO_AGENT_ROLE", "developer")
|
|
monkeypatch.setenv("ROBOCO_ORCHESTRATOR_URL", "http://test-orchestrator:8000")
|
|
monkeypatch.setenv("ROBOCO_SDK_URL", "http://test-sdk:9000")
|
|
monkeypatch.setenv("ROBOCO_TOOL_MANIFEST_PATH", str(manifest_path))
|
|
|
|
import roboco.mcp.flow_server as srv
|
|
|
|
importlib.reload(srv)
|
|
return srv
|
|
|
|
|
|
def _make_client(
|
|
orchestrator_response: dict[str, Any], sdk_response: dict[str, Any] | None
|
|
) -> Any:
|
|
"""Build an httpx.Client mock that dispatches by destination URL.
|
|
|
|
Calls hitting ``test-orchestrator`` return ``orchestrator_response``;
|
|
calls hitting ``test-sdk`` return ``sdk_response``. The mock also
|
|
records every URL+body it sees so assertions can verify the SDK was
|
|
(or wasn't) called.
|
|
"""
|
|
captured: list[tuple[str, dict[str, Any] | None]] = []
|
|
|
|
def _client_factory(*_args: Any, **_kwargs: Any) -> MagicMock:
|
|
client = MagicMock()
|
|
client.__enter__ = MagicMock(return_value=client)
|
|
client.__exit__ = MagicMock(return_value=False)
|
|
|
|
def _post(url: str, **kwargs: Any) -> MagicMock:
|
|
captured.append((url, kwargs.get("json")))
|
|
resp = MagicMock()
|
|
if "test-sdk" in url:
|
|
if sdk_response is None:
|
|
raise AssertionError("SDK called unexpectedly")
|
|
resp.json.return_value = sdk_response
|
|
else:
|
|
resp.json.return_value = orchestrator_response
|
|
return resp
|
|
|
|
client.post.side_effect = _post
|
|
return client
|
|
|
|
return _client_factory, captured
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Successful envelope — SDK must NOT be called
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_ok_envelope_does_not_touch_sdk(flow_module: types.ModuleType) -> None:
|
|
"""An envelope with error=None never POSTs to /verb/attempted."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"status": "awaiting_qa",
|
|
"task_id": "task-A",
|
|
"next": "wait",
|
|
"error": None,
|
|
},
|
|
sdk_response=None, # blow up if SDK is called
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A", notes="done")
|
|
|
|
assert result["status"] == "awaiting_qa"
|
|
assert result["error"] is None
|
|
# Only the orchestrator was contacted.
|
|
assert all("test-sdk" not in url for url, _ in captured)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Rejection envelope — SDK gets the verb + task_id + rejection_kind
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_rejection_forwards_to_sdk(flow_module: types.ModuleType) -> None:
|
|
"""A tracing_gap envelope triggers POST /verb/attempted with the right keys."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": "tracing_gap",
|
|
"missing": ["pr_number"],
|
|
"remediate": "open the PR",
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A", notes="done")
|
|
|
|
# Original rejection survives — breaker is not yet open.
|
|
assert result["error"] == "tracing_gap"
|
|
# Check the SDK saw the right payload.
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
sdk_url, sdk_body = sdk_calls[0]
|
|
assert sdk_url.endswith("/verb/attempted")
|
|
assert sdk_body == {
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"rejection_kind": "tracing_gap",
|
|
}
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"rejection_kind",
|
|
["tracing_gap", "invalid_state", "not_authorized", "incomplete_input"],
|
|
)
|
|
def test_all_counted_rejection_kinds_forwarded(
|
|
flow_module: types.ModuleType, rejection_kind: str
|
|
) -> None:
|
|
"""All four counted error kinds forward to the SDK."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": rejection_kind,
|
|
"message": "no",
|
|
"remediate": "fix",
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
flow_module.i_am_done("task-A")
|
|
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
_, sdk_body = sdk_calls[0]
|
|
assert sdk_body["rejection_kind"] == rejection_kind
|
|
|
|
|
|
def test_other_error_kinds_do_not_touch_sdk(flow_module: types.ModuleType) -> None:
|
|
"""not_found / transport_error aren't counted — SDK is not called."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={"error": "not_found", "message": "no task"},
|
|
sdk_response=None,
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
assert result["error"] == "not_found"
|
|
assert all("test-sdk" not in url for url, _ in captured)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Breaker open — envelope substituted with circuit_open
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_breaker_open_substitutes_envelope(flow_module: types.ModuleType) -> None:
|
|
"""When SDK reports open=true, the agent gets the circuit_open envelope."""
|
|
circuit_env: dict[str, Any] = {
|
|
"error": "circuit_open",
|
|
"message": ("verb 'i_am_done' rejected 3 times in last 60s — breaker open"),
|
|
"remediate": "call i_am_blocked or i_am_idle",
|
|
"context_briefing": {},
|
|
"status": None,
|
|
"task_id": None,
|
|
"next": None,
|
|
"evidence": {},
|
|
"correlation_id": None,
|
|
"current_state": None,
|
|
"valid_next_verbs": None,
|
|
}
|
|
factory, _ = _make_client(
|
|
orchestrator_response={"error": "tracing_gap", "remediate": "x"},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 3,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": True,
|
|
"circuit_envelope": circuit_env,
|
|
},
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
# The top-level envelope is the circuit_open substitution; the original
|
|
# fixable rejection is preserved nested as ``inner`` (#60).
|
|
assert result["error"] == "circuit_open"
|
|
assert "i_am_blocked" in result["remediate"]
|
|
assert result["inner"]["error"] == "tracing_gap"
|
|
assert result["inner"]["remediate"] == "x"
|
|
|
|
|
|
def test_breaker_open_preserves_original_as_inner(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""#60: the breaker substitution must not erase the fixable rejection.
|
|
|
|
The agent still needs the original kind/message/remediate to understand
|
|
WHY the verb failed — the circuit_open envelope only says the breaker
|
|
tripped. The original rejection is nested under ``inner``.
|
|
"""
|
|
circuit_env: dict[str, Any] = {
|
|
"error": "circuit_open",
|
|
"message": "verb 'i_am_done' rejected 3 times in 60s — breaker open",
|
|
"remediate": "call i_am_blocked or i_am_idle",
|
|
}
|
|
original = {
|
|
"error": "not_authorized",
|
|
"message": "you may not merge that PR",
|
|
"remediate": "PMs delegate; use delegate(...)",
|
|
"missing": [],
|
|
}
|
|
factory, _ = _make_client(
|
|
orchestrator_response=original,
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 3,
|
|
"limit": 3,
|
|
"open": True,
|
|
"circuit_envelope": circuit_env,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
assert result["error"] == "circuit_open"
|
|
# The original fixable rejection survives nested, not erased.
|
|
assert result["inner"] == original
|
|
# The SDK's envelope dict is not mutated in place (a fresh copy carries
|
|
# ``inner``) — important when the SDK reuses the envelope across calls.
|
|
assert "inner" not in circuit_env
|
|
|
|
|
|
def test_fourth_rejection_returns_circuit_open(flow_module: types.ModuleType) -> None:
|
|
"""Hitting the cap on the Nth call yields circuit_open on that same call.
|
|
|
|
The SDK records the attempt FIRST and reports open=true on the
|
|
response that just pushed it over the threshold — so the call that
|
|
trips the breaker is also the call that sees the substitution.
|
|
"""
|
|
# On the trip call the SDK reports open=true with the envelope.
|
|
circuit_env: dict[str, Any] = {
|
|
"error": "circuit_open",
|
|
"message": ("verb 'i_am_done' rejected 3 times in last 60s — breaker open"),
|
|
"remediate": "call i_am_blocked(reason='...') or i_am_idle()",
|
|
"context_briefing": {},
|
|
"status": None,
|
|
"task_id": None,
|
|
"next": None,
|
|
"evidence": {},
|
|
"correlation_id": None,
|
|
"current_state": None,
|
|
"valid_next_verbs": None,
|
|
}
|
|
factory, captured = _make_client(
|
|
orchestrator_response={"error": "tracing_gap", "remediate": "x"},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 3,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": True,
|
|
"circuit_envelope": circuit_env,
|
|
},
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
# Substituted envelope shown to the agent.
|
|
assert result["error"] == "circuit_open"
|
|
# Both calls happened: orchestrator (the verb) then SDK (record).
|
|
urls = [url for url, _ in captured]
|
|
assert any("test-orchestrator" in u for u in urls)
|
|
assert any("test-sdk" in u for u in urls)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Fail-open behaviour — SDK down must not break the gateway path
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_sdk_unreachable_fails_open(flow_module: types.ModuleType) -> None:
|
|
"""If the SDK raises, the agent still sees the original rejection."""
|
|
import httpx
|
|
|
|
def _client_factory(*_args: Any, **_kwargs: Any) -> MagicMock:
|
|
client = MagicMock()
|
|
client.__enter__ = MagicMock(return_value=client)
|
|
client.__exit__ = MagicMock(return_value=False)
|
|
|
|
def _post(url: str, **_kwargs: Any) -> MagicMock:
|
|
if "test-sdk" in url:
|
|
raise httpx.ConnectError("SDK down")
|
|
resp = MagicMock()
|
|
resp.json.return_value = {
|
|
"error": "tracing_gap",
|
|
"missing": ["pr_number"],
|
|
"remediate": "open the PR first",
|
|
}
|
|
return resp
|
|
|
|
client.post.side_effect = _post
|
|
return client
|
|
|
|
with patch("httpx.Client", side_effect=_client_factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
# SDK was unreachable — original envelope passes through.
|
|
assert result["error"] == "tracing_gap"
|
|
assert result["remediate"] == "open the PR first"
|
|
|
|
|
|
def test_sdk_returns_malformed_json_fails_open(flow_module: types.ModuleType) -> None:
|
|
"""If the SDK responds with un-JSON-able body, the original envelope wins."""
|
|
|
|
def _client_factory(*_args: Any, **_kwargs: Any) -> MagicMock:
|
|
client = MagicMock()
|
|
client.__enter__ = MagicMock(return_value=client)
|
|
client.__exit__ = MagicMock(return_value=False)
|
|
|
|
def _post(url: str, **_kwargs: Any) -> MagicMock:
|
|
resp = MagicMock()
|
|
if "test-sdk" in url:
|
|
resp.json.side_effect = ValueError("not json")
|
|
else:
|
|
resp.json.return_value = {
|
|
"error": "invalid_state",
|
|
"message": "wrong state",
|
|
"remediate": "transition first",
|
|
}
|
|
return resp
|
|
|
|
client.post.side_effect = _post
|
|
return client
|
|
|
|
with patch("httpx.Client", side_effect=_client_factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
|
|
assert result["error"] == "invalid_state"
|
|
assert result["message"] == "wrong state"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Verb extraction
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_verb_from_path_extracts_last_segment(flow_module: types.ModuleType) -> None:
|
|
"""_verb_from_path strips the role prefix and returns the verb token."""
|
|
assert (
|
|
flow_module._verb_from_path("/api/v1/flow/developer/i_am_done") == "i_am_done"
|
|
)
|
|
assert flow_module._verb_from_path("/api/v1/flow/qa/pass") == "pass"
|
|
assert (
|
|
flow_module._verb_from_path("/api/v1/flow/board/escalate_to_ceo")
|
|
== "escalate_to_ceo"
|
|
)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# task_id pass-through
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_task_id_none_is_forwarded_as_null(flow_module: types.ModuleType) -> None:
|
|
"""Verbs without a task_id (e.g. give_me_work) post task_id=None to the SDK."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={"error": "tracing_gap", "remediate": "x"},
|
|
sdk_response={
|
|
"verb": "give_me_work",
|
|
"task_id": None,
|
|
"attempts": 1,
|
|
"limit": None, # unlimited verb
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
|
|
with patch("httpx.Client", side_effect=factory):
|
|
flow_module.give_me_work()
|
|
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
_, sdk_body = sdk_calls[0]
|
|
assert sdk_body["task_id"] is None
|
|
assert sdk_body["verb"] == "give_me_work"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Non-string-error / no-error-field rejection shapes are counted
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_422_validation_failure_counts_as_incomplete_input(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A 422 validation-failure body (`{"detail": [...]}`, no `error`) must count
|
|
toward the breaker as `incomplete_input` — a storm of 422s is retry-storm-worthy.
|
|
Mirrors do_server's classifier (the two servers share the same breaker logic).
|
|
"""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"detail": [
|
|
{
|
|
"loc": ["body", "task_id"],
|
|
"msg": "field required",
|
|
"type": "value_error.missing",
|
|
}
|
|
],
|
|
"body": None,
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
flow_module.i_am_done("task-A")
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
assert sdk_calls[0][1]["rejection_kind"] == "incomplete_input"
|
|
|
|
|
|
def test_dict_shaped_internal_error_counts_as_invalid_state(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A 500 INTERNAL_ERROR dict-shaped response (generic_exception_handler) counts
|
|
as `invalid_state` — a storm of 500s is retry-storm-worthy.
|
|
"""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": {
|
|
"code": "INTERNAL_ERROR",
|
|
"message": "An internal error occurred",
|
|
"details": {"correlation_id": "abc"},
|
|
}
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
flow_module.i_am_done("task-A")
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
assert sdk_calls[0][1]["rejection_kind"] == "invalid_state"
|
|
|
|
|
|
def test_dict_shaped_not_found_normalized_to_envelope_and_not_counted(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A dict-shaped NOT_FOUND (404 family) is normalized to an Envelope
|
|
(`error="not_found"` string + lifted message + remediate) — the agent
|
|
never sees a dict `error`, which violates the Envelope wire format — and
|
|
the breaker is NOT touched (parity with the string `not_found` contract).
|
|
"""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": {"code": "TASK_NOT_FOUND", "message": "no such task"}
|
|
},
|
|
sdk_response=None, # SDK must not be touched
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
# Normalized to a string-kind Envelope — the dict `error` never reaches
|
|
# the agent (#232).
|
|
assert result["error"] == "not_found"
|
|
assert result["message"] == "no such task"
|
|
assert isinstance(result["remediate"], str) and result["remediate"]
|
|
assert result["missing"] == []
|
|
# NOT_FOUND still does not count toward the breaker.
|
|
assert all("test-sdk" not in url for url, _ in captured)
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# A manifest-registered verb whose route is missing must return an envelope
|
|
# rejection (not a raw 404 body) so the breaker counts it.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def _make_404_client() -> tuple[Any, list[tuple[str, dict[str, Any] | None]]]:
|
|
"""Build an httpx.Client mock whose orchestrator call returns FastAPI's
|
|
default 404 body (``{"detail": "Not Found"}``, status 404) — the shape a
|
|
manifest-registered verb sees when its route is missing. The SDK call
|
|
returns a not-yet-open breaker so the original envelope is preserved.
|
|
"""
|
|
captured: list[tuple[str, dict[str, Any] | None]] = []
|
|
|
|
def _client_factory(*_args: Any, **_kwargs: Any) -> MagicMock:
|
|
client = MagicMock()
|
|
client.__enter__ = MagicMock(return_value=client)
|
|
client.__exit__ = MagicMock(return_value=False)
|
|
|
|
def _post(url: str, **kwargs: Any) -> MagicMock:
|
|
captured.append((url, kwargs.get("json")))
|
|
resp = MagicMock()
|
|
if "test-sdk" in url:
|
|
resp.json.return_value = {
|
|
"verb": "triage",
|
|
"task_id": None,
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
}
|
|
else:
|
|
# FastAPI's default 404 for a missing route.
|
|
resp.status_code = 404
|
|
resp.json.return_value = {"detail": "Not Found"}
|
|
return resp
|
|
|
|
client.post.side_effect = _post
|
|
return client
|
|
|
|
return _client_factory, captured
|
|
|
|
|
|
def test_missing_route_404_returns_envelope_and_counts(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A 404 from the orchestrator (manifest-registered verb with no route) must
|
|
surface as a proper `invalid_state` Envelope rejection — not FastAPI's raw
|
|
``{"detail": "Not Found"}`` body — and the breaker must count it.
|
|
"""
|
|
factory, captured = _make_404_client()
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.triage()
|
|
# Envelope rejection, not the raw 404 body.
|
|
assert result["error"] == "invalid_state"
|
|
assert "remediate" in result
|
|
assert "detail" not in result # the raw 404 body was not passed through
|
|
# Breaker was notified so a storm of these trips it.
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
_, sdk_body = sdk_calls[0]
|
|
assert sdk_body is not None
|
|
assert sdk_body["rejection_kind"] == "invalid_state"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# #232: a non-404 JSON exception-handler body (dict `error` / 422 `detail`)
|
|
# must be normalized to an Envelope wire format before reaching the agent.
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_exception_dict_500_normalized_to_invalid_state_envelope(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A 500 with a dict-shaped error (generic_exception_handler) is normalized
|
|
to an Envelope: top-level `error='invalid_state'` (string kind), the dict's
|
|
`message` lifted to top-level `message`, a non-empty `remediate`, `missing=[]`.
|
|
The agent never receives a dict `error` (which violates the Envelope contract
|
|
it's prompted to trust). The breaker still counts it as `invalid_state`."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": {
|
|
"code": "INTERNAL_ERROR",
|
|
"message": "An internal error occurred",
|
|
"details": {"correlation_id": "abc"},
|
|
}
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
assert result["error"] == "invalid_state"
|
|
assert result["message"] == "An internal error occurred"
|
|
assert isinstance(result["remediate"], str) and result["remediate"]
|
|
assert result["missing"] == []
|
|
# The synthesized Envelope still flows through the breaker (string kind in
|
|
# the counted set) so a 500 storm trips it.
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
assert sdk_calls[0][1]["rejection_kind"] == "invalid_state"
|
|
|
|
|
|
def test_422_detail_normalized_to_incomplete_input_envelope(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A 422 request-validation body (`{"detail": [...]}`, no `error`) is
|
|
normalized to an Envelope with `error='incomplete_input'`, a non-empty
|
|
`remediate`, and `missing=[]`; the raw validation `detail` is preserved so
|
|
the agent can see WHICH fields failed. The breaker still counts it."""
|
|
detail_body = [
|
|
{"loc": ["body", "task_id"], "msg": "field required", "type": "missing"}
|
|
]
|
|
factory, captured = _make_client(
|
|
orchestrator_response={"detail": detail_body, "body": None},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
assert result["error"] == "incomplete_input"
|
|
assert isinstance(result["remediate"], str) and result["remediate"]
|
|
assert result["missing"] == []
|
|
# The validation detail survives so the agent knows which fields to fix.
|
|
assert result["detail"] == detail_body
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
assert sdk_calls[0][1]["rejection_kind"] == "incomplete_input"
|
|
|
|
|
|
def test_dict_authorized_normalized_to_not_authorized_envelope(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""A dict-shaped PERMISSION_DENIED maps to `error='not_authorized'` in the
|
|
normalized Envelope (exact-code map, not a substring accident — #161), with
|
|
the message lifted and a remediate. The breaker counts it as not_authorized."""
|
|
factory, captured = _make_client(
|
|
orchestrator_response={
|
|
"error": {
|
|
"code": "PERMISSION_DENIED",
|
|
"message": "you may not merge that PR",
|
|
"details": {},
|
|
}
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "task-A",
|
|
"attempts": 1,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": False,
|
|
"circuit_envelope": None,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
assert result["error"] == "not_authorized"
|
|
assert result["message"] == "you may not merge that PR"
|
|
assert isinstance(result["remediate"], str) and result["remediate"]
|
|
assert result["missing"] == []
|
|
sdk_calls = [(url, body) for url, body in captured if "test-sdk" in url]
|
|
assert len(sdk_calls) == 1
|
|
assert sdk_calls[0][1]["rejection_kind"] == "not_authorized"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# #359: the circuit_open substitution lifts task_id/correlation_id from the
|
|
# original rejection to the top-level envelope (the SDK's envelope omits them).
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_circuit_open_lifts_task_id_and_correlation_id(
|
|
flow_module: types.ModuleType,
|
|
) -> None:
|
|
"""When the breaker trips, the SDK's circuit_envelope carries no task_id /
|
|
correlation_id, but the original rejection payload does (the orchestrator
|
|
stamps them on every envelope). The substitution must lift them to the
|
|
top-level so the agent's envelope contract (read top-level task_id /
|
|
correlation_id) and ops audit-join still work — not leave them only nested
|
|
in `inner`."""
|
|
circuit_env: dict[str, Any] = {
|
|
"error": "circuit_open",
|
|
"message": "verb 'i_am_done' rejected 3 times in 60s — breaker open",
|
|
"remediate": "call i_am_blocked or i_am_idle",
|
|
}
|
|
factory, _ = _make_client(
|
|
orchestrator_response={
|
|
"error": "tracing_gap",
|
|
"remediate": "open the PR",
|
|
"task_id": "T1",
|
|
"correlation_id": "C1",
|
|
},
|
|
sdk_response={
|
|
"verb": "i_am_done",
|
|
"task_id": "T1",
|
|
"attempts": 3,
|
|
"limit": 3,
|
|
"window_seconds": 60,
|
|
"open": True,
|
|
"circuit_envelope": circuit_env,
|
|
},
|
|
)
|
|
with patch("httpx.Client", side_effect=factory):
|
|
result = flow_module.i_am_done("task-A")
|
|
assert result["error"] == "circuit_open"
|
|
# Lifted from the original rejection to the top level (not just in inner).
|
|
assert result["task_id"] == "T1"
|
|
assert result["correlation_id"] == "C1"
|
|
# The original rejection still survives nested.
|
|
assert result["inner"]["error"] == "tracing_gap"
|
|
assert result["inner"]["task_id"] == "T1"
|