fix(agent): allow remaining actions for an approved task (#6113)

* fix(agent): allow remaining actions for an approved task

* fix(agent): make approval continuation control-only

* fix(ci): preserve approval taint and cache-buster contract

* fix(ui): keep tool approvals in current chat

* fix(ui): route tool approvals through chat submit

* test(ui): pin approval submit routing

* fix(agent): complete approval denial flow

* fix(ui): avoid duplicate ask-user close icon

* fix(agent): retain approved tool in continuation set

* revert(ui): keep PR 6113 scoped to approval continuation

* fix(agent): add task and chat approval scopes

* fix(ui): prevent duplicate ask-user close icon

* feat(ui): add ask-user option shortcuts

* fix(compare): route ask-user choices per pane

* fix(agent): keep skill-test approvals to a single action

The chat card now reuses the wire value `approve` to mean chat-session
scope, and `consume()` returned `allow_remaining_actions=True` for it
unconditionally. The skill-test approval route was never updated: it still
sends `approve` meaning "once", and its button still reads "Allow once",
but the grant it got back set `approval_gate_bypassed` for the rest of the
resumed run. That surface wraps the skill body and every transcript byte
as untrusted context, so it is the last place where one click should
ungate everything that follows.

Give `consume()` an explicit `allow_continuation` flag. Callers that own a
resumable chat keep the scope the user picked; callers that do not — the
skill tester, unattended audits — get SINGLE_ACTION and the gate re-arms
behind the sealed action, which is what their label promises.

* fix(ui): cache-bust every module the approval click depends on

chatStream.js, compare/index.js and compare/stream.js all changed
behaviour but kept their old `?v=`, while chat.js and chatRenderer.js were
bumped. A returning browser therefore serves the new chat.js — which now
deliberately leaves the composer empty and clicks the send button — next to
the cached chatStream.js that has no interceptor. With an empty composer
that button sits at `data-mode="newchat"`, so the click opens a new chat
and the approval is dropped.

Bump the three, and version compare/stream.js's chatRenderer import to
match everyone else's so the ask_user keydown listener binds to one module
instance instead of two.

* fix(ui): keep the digit shortcuts off tool approval cards

With an approval card on screen and focus anywhere outside an input, a bare
`1` fired `approve_task` — the widest of the three grants — with no
modifier and no confirmation. That card is the one control whose entire
purpose is deliberate consent after untrusted context influenced the run,
and Deny sits at 3.

Label the card with its kind and skip the shortcut for approvals. Ordinary
ask_user questions keep 1-3.

* fix(compare): restore a pane's ask_user card instead of dropping the choice

renderAskUserCard removes the card as soon as onSubmit accepts, but the
resume loop gave up silently after 10s if the originating stream still owned
the pane. The user saw the click land, the card vanish, and nothing happen,
with no way to get it back.

Re-render the card on that deadline and say why. The reroll case still
returns without sending — that choice belongs to a stream that no longer
exists.

* refactor(chat): drop the unreachable deny branch

`if decision != "deny"` is always true — the deny path returns a
StreamingResponse a few lines above. It reads as if deny still falls
through to the toggle restore.

---------

Co-authored-by: Léo <leograndcontact@gmail.com>
This commit is contained in:
RaresKeY
2026-08-19 08:01:34 -06:00
committed by GitHub
co-authored by Léo
parent 5c835014ac
commit 981652358e
20 changed files with 1361 additions and 119 deletions
+51 -1
View File
@@ -8,6 +8,11 @@ These are simple datacontainers. All persistence is handled by SessionManager.
from dataclasses import dataclass
from typing import Dict, List, Any, Optional, TYPE_CHECKING
from src.tool_approval_scopes import (
CHAT_SESSION_APPROVAL_CONTEXT_MARKER,
CHAT_SESSION_APPROVAL_DECISION,
)
if TYPE_CHECKING:
from .session_manager import SessionManager
@@ -31,6 +36,35 @@ set_session_manager = set_session_manager_instance
get_session_manager = get_session_manager_instance
def _history_grants_chat_session_approval(
history: List["ChatMessage"],
session_id: str,
) -> bool:
"""Return whether this exact chat has a resolved session-scope grant."""
expected_session = str(session_id or "")
if not expected_session:
return False
for message in reversed(history or []):
metadata = getattr(message, "metadata", None)
if not isinstance(metadata, dict):
continue
tool_events = metadata.get("tool_events")
if not isinstance(tool_events, list):
continue
for event in reversed(tool_events):
ask_user = event.get("ask_user") if isinstance(event, dict) else None
if not isinstance(ask_user, dict):
continue
if (
ask_user.get("kind") == "tool_approval"
and ask_user.get("resolved") == CHAT_SESSION_APPROVAL_DECISION
and str(ask_user.get("session_id") or "") == expected_session
):
return True
return False
@dataclass
class ChatMessage:
"""A single chat message."""
@@ -116,11 +150,27 @@ class Session:
the model. Display/history-load paths use the raw ``history`` and are
unaffected.
"""
return [
messages = [
msg.to_dict()
for msg in self.history
if (msg.metadata or {}).get("source") != "slash"
]
if not _history_grants_chat_session_approval(self.history, self.id):
return messages
# Keep the grant close to the latest user request so route-neutral
# compaction/trimming preserves it. Copy the metadata instead of
# mutating the durable transcript object.
for index in range(len(messages) - 1, -1, -1):
if messages[index].get("role") != "user":
continue
message = dict(messages[index])
metadata = dict(message.get("metadata") or {})
metadata[CHAT_SESSION_APPROVAL_CONTEXT_MARKER] = True
message["metadata"] = metadata
messages[index] = message
break
return messages
def get(self, key: str, default=None):
"""Dict-like access for compatibility."""