Files
hermes-agent/gateway/session.py

1373 lines
54 KiB
Python
Raw Normal View History

"""
Session management for the gateway.
Handles:
- Session context tracking (where messages come from)
- Session storage (conversations persisted to disk)
- Reset policy evaluation (when to start fresh)
- Dynamic system prompt injection (agent knows its context)
"""
import hashlib
import logging
import os
import json
import threading
import uuid
from pathlib import Path
from datetime import datetime, timedelta
from dataclasses import dataclass
from typing import Dict, List, Optional, Any
logger = logging.getLogger(__name__)
def _now() -> datetime:
"""Return the current local time."""
return datetime.now()
# ---------------------------------------------------------------------------
# PII redaction helpers
# ---------------------------------------------------------------------------
def _hash_id(value: str) -> str:
"""Deterministic 12-char hex hash of an identifier."""
return hashlib.sha256(value.encode("utf-8")).hexdigest()[:12]
def _hash_sender_id(value: str) -> str:
"""Hash a sender ID to ``user_<12hex>``."""
return f"user_{_hash_id(value)}"
def _hash_chat_id(value: str) -> str:
"""Hash the numeric portion of a chat ID, preserving platform prefix.
``telegram:12345`` ``telegram:<hash>``
``12345`` ``<hash>``
"""
colon = value.find(":")
if colon > 0:
prefix = value[:colon]
return f"{prefix}:{_hash_id(value[colon + 1:])}"
return _hash_id(value)
from .config import (
Platform,
GatewayConfig,
SessionResetPolicy, # noqa: F401 — re-exported via gateway/__init__.py
HomeChannel,
)
from .whatsapp_identity import (
canonical_whatsapp_identifier,
normalize_whatsapp_identifier,
)
@dataclass
class SessionSource:
"""
Describes where a message originated from.
This information is used to:
1. Route responses back to the right place
2. Inject context into the system prompt
3. Track origin for cron job delivery
"""
platform: Platform
chat_id: str
chat_name: Optional[str] = None
chat_type: str = "dm" # "dm", "group", "channel", "thread"
user_id: Optional[str] = None
user_name: Optional[str] = None
thread_id: Optional[str] = None # For forum topics, Discord threads, etc.
chat_topic: Optional[str] = None # Channel topic/description (Discord, Slack)
user_id_alt: Optional[str] = None # Platform-specific stable alt ID (Signal UUID, Feishu union_id)
feat: add Signal messenger gateway platform (#405) Complete Signal adapter using signal-cli daemon HTTP API. Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes. Architecture: - SSE streaming for inbound messages with exponential backoff (2s→60s) - JSON-RPC 2.0 for outbound (send, typing, attachments, contacts) - Health monitor detects stale SSE connections (120s threshold) - Phone number redaction in all logs and global redact.py Features: - DM and group message support with separate access policies - DM policies: pairing (default), allowlist, open - Group policies: disabled (default), allowlist, open - Attachment download with magic-byte type detection - Typing indicators (8s refresh interval) - 100MB attachment size limit, 8000 char message limit - E.164 phone + UUID allowlist support Integration: - Platform.SIGNAL enum in gateway/config.py - Signal in _is_user_authorized() allowlist maps (gateway/run.py) - Adapter factory in _create_adapter() (gateway/run.py) - user_id_alt/chat_id_alt fields in SessionSource for UUIDs - send_message tool support via httpx JSON-RPC (not aiohttp) - Interactive setup wizard in 'hermes gateway setup' - Connectivity testing during setup (pings /api/v1/check) - signal-cli detection and install guidance Bug fixes from PR #268: - Timestamp reads from envelope_data (not outer wrapper) - Uses httpx consistently (not aiohttp in send_message tool) - SIGNAL_DEBUG scoped to signal logger (not root) - extract_images regex NOT modified (preserves group numbering) - pairing.py NOT modified (no cross-platform side effects) - No dual authorization (adapter defers to run.py for user auth) - Wildcard uses set membership ('*' in set, not list equality) - .zip default for PK magic bytes (not .docx) No new Python dependencies — uses httpx (already core). External requirement: signal-cli daemon (user-installed). Tests: 30 new tests covering config, init, helpers, session source, phone redaction, authorization, and send_message integration. Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
2026-03-08 20:20:35 -07:00
chat_id_alt: Optional[str] = None # Signal group internal ID
fix(discord): DISCORD_ALLOW_BOTS=mentions/all now works without DISCORD_ALLOWED_USERS Fixes #4466. Root cause: two sequential authorization gates both independently rejected bot messages, making DISCORD_ALLOW_BOTS completely ineffective. Gate 1 — `discord.py` `on_message`: _is_allowed_user ran BEFORE the bot filter, so bot senders were dropped before the DISCORD_ALLOW_BOTS policy was ever evaluated. Gate 2 — `gateway/run.py` _is_user_authorized: The gateway-level allowlist check rejected bot IDs with 'Unauthorized user: <bot_id>' even if they passed Gate 1. Fix: gateway/platforms/discord.py — reorder on_message so DISCORD_ALLOW_BOTS runs BEFORE _is_allowed_user. Bots permitted by the filter skip the user allowlist; non-bots are still checked. gateway/session.py — add is_bot: bool = False to SessionSource so the gateway layer can distinguish bot senders. gateway/platforms/base.py — expose is_bot parameter in build_source. gateway/platforms/discord.py _handle_message — set is_bot=True when building the SessionSource for bot authors. gateway/run.py _is_user_authorized — when source.is_bot is True AND DISCORD_ALLOW_BOTS is 'mentions' or 'all', return True early. Platform filter already validated the message at on_message; don't re-reject. Behavior matrix: | Config | Before | After | | DISCORD_ALLOW_BOTS=none (default) | Blocked | Blocked | | DISCORD_ALLOW_BOTS=all | Blocked | Allowed | | DISCORD_ALLOW_BOTS=mentions + @mention | Blocked | Allowed | | DISCORD_ALLOW_BOTS=mentions, no mention | Blocked | Blocked | | Human in DISCORD_ALLOWED_USERS | Allowed | Allowed | | Human NOT in DISCORD_ALLOWED_USERS | Blocked | Blocked | Co-authored-by: Hermes Maintainer <hermes@nousresearch.com>
2026-04-17 05:41:19 -07:00
is_bot: bool = False # True when the message author is a bot/webhook (Discord)
guild_id: Optional[str] = None # Discord guild / Slack workspace / Matrix server scope
parent_chat_id: Optional[str] = None # Parent channel when chat_id refers to a thread
message_id: Optional[str] = None # ID of the triggering message (for pin/reply/react)
@property
def description(self) -> str:
"""Human-readable description of the source."""
if self.platform == Platform.LOCAL:
return "CLI terminal"
parts = []
if self.chat_type == "dm":
parts.append(f"DM with {self.user_name or self.user_id or 'user'}")
elif self.chat_type == "group":
parts.append(f"group: {self.chat_name or self.chat_id}")
elif self.chat_type == "channel":
parts.append(f"channel: {self.chat_name or self.chat_id}")
else:
parts.append(self.chat_name or self.chat_id)
if self.thread_id:
parts.append(f"thread: {self.thread_id}")
return ", ".join(parts)
def to_dict(self) -> Dict[str, Any]:
feat: add Signal messenger gateway platform (#405) Complete Signal adapter using signal-cli daemon HTTP API. Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes. Architecture: - SSE streaming for inbound messages with exponential backoff (2s→60s) - JSON-RPC 2.0 for outbound (send, typing, attachments, contacts) - Health monitor detects stale SSE connections (120s threshold) - Phone number redaction in all logs and global redact.py Features: - DM and group message support with separate access policies - DM policies: pairing (default), allowlist, open - Group policies: disabled (default), allowlist, open - Attachment download with magic-byte type detection - Typing indicators (8s refresh interval) - 100MB attachment size limit, 8000 char message limit - E.164 phone + UUID allowlist support Integration: - Platform.SIGNAL enum in gateway/config.py - Signal in _is_user_authorized() allowlist maps (gateway/run.py) - Adapter factory in _create_adapter() (gateway/run.py) - user_id_alt/chat_id_alt fields in SessionSource for UUIDs - send_message tool support via httpx JSON-RPC (not aiohttp) - Interactive setup wizard in 'hermes gateway setup' - Connectivity testing during setup (pings /api/v1/check) - signal-cli detection and install guidance Bug fixes from PR #268: - Timestamp reads from envelope_data (not outer wrapper) - Uses httpx consistently (not aiohttp in send_message tool) - SIGNAL_DEBUG scoped to signal logger (not root) - extract_images regex NOT modified (preserves group numbering) - pairing.py NOT modified (no cross-platform side effects) - No dual authorization (adapter defers to run.py for user auth) - Wildcard uses set membership ('*' in set, not list equality) - .zip default for PK magic bytes (not .docx) No new Python dependencies — uses httpx (already core). External requirement: signal-cli daemon (user-installed). Tests: 30 new tests covering config, init, helpers, session source, phone redaction, authorization, and send_message integration. Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
2026-03-08 20:20:35 -07:00
d = {
"platform": self.platform.value,
"chat_id": self.chat_id,
"chat_name": self.chat_name,
"chat_type": self.chat_type,
"user_id": self.user_id,
"user_name": self.user_name,
"thread_id": self.thread_id,
"chat_topic": self.chat_topic,
}
feat: add Signal messenger gateway platform (#405) Complete Signal adapter using signal-cli daemon HTTP API. Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes. Architecture: - SSE streaming for inbound messages with exponential backoff (2s→60s) - JSON-RPC 2.0 for outbound (send, typing, attachments, contacts) - Health monitor detects stale SSE connections (120s threshold) - Phone number redaction in all logs and global redact.py Features: - DM and group message support with separate access policies - DM policies: pairing (default), allowlist, open - Group policies: disabled (default), allowlist, open - Attachment download with magic-byte type detection - Typing indicators (8s refresh interval) - 100MB attachment size limit, 8000 char message limit - E.164 phone + UUID allowlist support Integration: - Platform.SIGNAL enum in gateway/config.py - Signal in _is_user_authorized() allowlist maps (gateway/run.py) - Adapter factory in _create_adapter() (gateway/run.py) - user_id_alt/chat_id_alt fields in SessionSource for UUIDs - send_message tool support via httpx JSON-RPC (not aiohttp) - Interactive setup wizard in 'hermes gateway setup' - Connectivity testing during setup (pings /api/v1/check) - signal-cli detection and install guidance Bug fixes from PR #268: - Timestamp reads from envelope_data (not outer wrapper) - Uses httpx consistently (not aiohttp in send_message tool) - SIGNAL_DEBUG scoped to signal logger (not root) - extract_images regex NOT modified (preserves group numbering) - pairing.py NOT modified (no cross-platform side effects) - No dual authorization (adapter defers to run.py for user auth) - Wildcard uses set membership ('*' in set, not list equality) - .zip default for PK magic bytes (not .docx) No new Python dependencies — uses httpx (already core). External requirement: signal-cli daemon (user-installed). Tests: 30 new tests covering config, init, helpers, session source, phone redaction, authorization, and send_message integration. Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
2026-03-08 20:20:35 -07:00
if self.user_id_alt:
d["user_id_alt"] = self.user_id_alt
if self.chat_id_alt:
d["chat_id_alt"] = self.chat_id_alt
if self.guild_id:
d["guild_id"] = self.guild_id
if self.parent_chat_id:
d["parent_chat_id"] = self.parent_chat_id
if self.message_id:
d["message_id"] = self.message_id
feat: add Signal messenger gateway platform (#405) Complete Signal adapter using signal-cli daemon HTTP API. Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes. Architecture: - SSE streaming for inbound messages with exponential backoff (2s→60s) - JSON-RPC 2.0 for outbound (send, typing, attachments, contacts) - Health monitor detects stale SSE connections (120s threshold) - Phone number redaction in all logs and global redact.py Features: - DM and group message support with separate access policies - DM policies: pairing (default), allowlist, open - Group policies: disabled (default), allowlist, open - Attachment download with magic-byte type detection - Typing indicators (8s refresh interval) - 100MB attachment size limit, 8000 char message limit - E.164 phone + UUID allowlist support Integration: - Platform.SIGNAL enum in gateway/config.py - Signal in _is_user_authorized() allowlist maps (gateway/run.py) - Adapter factory in _create_adapter() (gateway/run.py) - user_id_alt/chat_id_alt fields in SessionSource for UUIDs - send_message tool support via httpx JSON-RPC (not aiohttp) - Interactive setup wizard in 'hermes gateway setup' - Connectivity testing during setup (pings /api/v1/check) - signal-cli detection and install guidance Bug fixes from PR #268: - Timestamp reads from envelope_data (not outer wrapper) - Uses httpx consistently (not aiohttp in send_message tool) - SIGNAL_DEBUG scoped to signal logger (not root) - extract_images regex NOT modified (preserves group numbering) - pairing.py NOT modified (no cross-platform side effects) - No dual authorization (adapter defers to run.py for user auth) - Wildcard uses set membership ('*' in set, not list equality) - .zip default for PK magic bytes (not .docx) No new Python dependencies — uses httpx (already core). External requirement: signal-cli daemon (user-installed). Tests: 30 new tests covering config, init, helpers, session source, phone redaction, authorization, and send_message integration. Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
2026-03-08 20:20:35 -07:00
return d
@classmethod
def from_dict(cls, data: Dict[str, Any]) -> "SessionSource":
return cls(
platform=Platform(data["platform"]),
chat_id=str(data["chat_id"]),
chat_name=data.get("chat_name"),
chat_type=data.get("chat_type", "dm"),
user_id=data.get("user_id"),
user_name=data.get("user_name"),
thread_id=data.get("thread_id"),
chat_topic=data.get("chat_topic"),
feat: add Signal messenger gateway platform (#405) Complete Signal adapter using signal-cli daemon HTTP API. Based on PR #268 by ibhagwan, rebuilt on current main with bug fixes. Architecture: - SSE streaming for inbound messages with exponential backoff (2s→60s) - JSON-RPC 2.0 for outbound (send, typing, attachments, contacts) - Health monitor detects stale SSE connections (120s threshold) - Phone number redaction in all logs and global redact.py Features: - DM and group message support with separate access policies - DM policies: pairing (default), allowlist, open - Group policies: disabled (default), allowlist, open - Attachment download with magic-byte type detection - Typing indicators (8s refresh interval) - 100MB attachment size limit, 8000 char message limit - E.164 phone + UUID allowlist support Integration: - Platform.SIGNAL enum in gateway/config.py - Signal in _is_user_authorized() allowlist maps (gateway/run.py) - Adapter factory in _create_adapter() (gateway/run.py) - user_id_alt/chat_id_alt fields in SessionSource for UUIDs - send_message tool support via httpx JSON-RPC (not aiohttp) - Interactive setup wizard in 'hermes gateway setup' - Connectivity testing during setup (pings /api/v1/check) - signal-cli detection and install guidance Bug fixes from PR #268: - Timestamp reads from envelope_data (not outer wrapper) - Uses httpx consistently (not aiohttp in send_message tool) - SIGNAL_DEBUG scoped to signal logger (not root) - extract_images regex NOT modified (preserves group numbering) - pairing.py NOT modified (no cross-platform side effects) - No dual authorization (adapter defers to run.py for user auth) - Wildcard uses set membership ('*' in set, not list equality) - .zip default for PK magic bytes (not .docx) No new Python dependencies — uses httpx (already core). External requirement: signal-cli daemon (user-installed). Tests: 30 new tests covering config, init, helpers, session source, phone redaction, authorization, and send_message integration. Co-authored-by: ibhagwan <ibhagwan@users.noreply.github.com>
2026-03-08 20:20:35 -07:00
user_id_alt=data.get("user_id_alt"),
chat_id_alt=data.get("chat_id_alt"),
guild_id=data.get("guild_id"),
parent_chat_id=data.get("parent_chat_id"),
message_id=data.get("message_id"),
)
@dataclass
class SessionContext:
"""
Full context for a session, used for dynamic system prompt injection.
The agent receives this information to understand:
- Where messages are coming from
- What platforms are available
- Where it can deliver scheduled task outputs
"""
source: SessionSource
connected_platforms: List[Platform]
home_channels: Dict[Platform, HomeChannel]
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
shared_multi_user_session: bool = False
# Session metadata
session_key: str = ""
session_id: str = ""
created_at: Optional[datetime] = None
updated_at: Optional[datetime] = None
def to_dict(self) -> Dict[str, Any]:
return {
"source": self.source.to_dict(),
"connected_platforms": [p.value for p in self.connected_platforms],
"home_channels": {
p.value: hc.to_dict() for p, hc in self.home_channels.items()
},
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
"shared_multi_user_session": self.shared_multi_user_session,
"session_key": self.session_key,
"session_id": self.session_id,
"created_at": self.created_at.isoformat() if self.created_at else None,
"updated_at": self.updated_at.isoformat() if self.updated_at else None,
}
_PII_SAFE_PLATFORMS = frozenset({
Platform.WHATSAPP,
Platform.SIGNAL,
Platform.TELEGRAM,
Platform.BLUEBUBBLES,
})
"""Platforms where user IDs can be safely redacted (no in-message mention system
that requires raw IDs). Discord is excluded because mentions use ``<@user_id>``
and the LLM needs the real ID to tag users."""
def _discord_tools_loaded() -> bool:
"""True iff the agent will actually have Discord tools this session.
Two conditions must hold:
1. The `discord` or `discord_admin` toolset is enabled for the
Discord platform via `hermes tools` (opt-in, default OFF).
2. `DISCORD_BOT_TOKEN` is set the tool's `check_fn` gates on it
at registry time, so the toolset being enabled in config is not
enough if the token isn't configured.
Returns False (safe default keeps the stale-API disclaimer) on any
error so a bad config can't silently promise tools the agent lacks.
"""
if not (os.environ.get("DISCORD_BOT_TOKEN") or "").strip():
return False
try:
from hermes_cli.config import load_config
from hermes_cli.tools_config import _get_platform_tools
cfg = load_config()
enabled = _get_platform_tools(cfg, "discord", include_default_mcp_servers=False)
return "discord" in enabled or "discord_admin" in enabled
except Exception:
return False
def build_session_context_prompt(
context: SessionContext,
*,
redact_pii: bool = False,
) -> str:
"""
Build the dynamic system prompt section that tells the agent about its context.
This is injected into the system prompt so the agent knows:
- Where messages are coming from
- What platforms are connected
- Where it can deliver scheduled task outputs
When *redact_pii* is True **and** the source platform is in
``_PII_SAFE_PLATFORMS``, phone numbers are stripped and user/chat IDs
are replaced with deterministic hashes before being sent to the LLM.
Platforms like Discord are excluded because mentions need real IDs.
Routing still uses the original values (they stay in SessionSource).
"""
# Only apply redaction on platforms where IDs aren't needed for mentions
redact_pii = redact_pii and context.source.platform in _PII_SAFE_PLATFORMS
lines = [
"## Current Session Context",
"",
]
# Source info
platform_name = context.source.platform.value.title()
if context.source.platform == Platform.LOCAL:
lines.append(f"**Source:** {platform_name} (the machine running this agent)")
else:
# Build a description that respects PII redaction
src = context.source
if redact_pii:
# Build a safe description without raw IDs
_uname = src.user_name or (
_hash_sender_id(src.user_id) if src.user_id else "user"
)
_cname = src.chat_name or _hash_chat_id(src.chat_id)
if src.chat_type == "dm":
desc = f"DM with {_uname}"
elif src.chat_type == "group":
desc = f"group: {_cname}"
elif src.chat_type == "channel":
desc = f"channel: {_cname}"
else:
desc = _cname
else:
desc = src.description
lines.append(f"**Source:** {platform_name} ({desc})")
# Channel topic (if available - provides context about the channel's purpose)
if context.source.chat_topic:
lines.append(f"**Channel Topic:** {context.source.chat_topic}")
# User identity.
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
# In shared multi-user sessions (shared threads OR shared non-thread groups
# when group_sessions_per_user=False), multiple users contribute to the same
# conversation. Don't pin a single user name in the system prompt — it
# changes per-turn and would bust the prompt cache. Instead, note that
# this is a multi-user session; individual sender names are prefixed on
# each user message by the gateway.
if context.shared_multi_user_session:
session_label = "Multi-user thread" if context.source.thread_id else "Multi-user session"
lines.append(
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
f"**Session type:** {session_label} — messages are prefixed "
"with [sender name]. Multiple users may participate."
)
elif context.source.user_name:
lines.append(f"**User:** {context.source.user_name}")
elif context.source.user_id:
uid = context.source.user_id
if redact_pii:
uid = _hash_sender_id(uid)
lines.append(f"**User ID:** {uid}")
# Platform-specific behavioral notes
if context.source.platform == Platform.SLACK:
lines.append("")
lines.append(
"**Platform notes:** You are running inside Slack. "
"You do NOT have access to Slack-specific APIs — you cannot search "
"channel history, pin/unpin messages, manage channels, or list users. "
fix(slack): extract rich_text quotes/lists and link unfurl previews Slack's modern composer sends messages with a 'blocks' array that contains rich_text elements. When a user forwards or quotes another message, the quoted content shows up in the rich_text_quote children of that array — and is NOT included in the plain 'text' field. The agent saw only the lossy plain text and was blind to forwarded / quoted content. Same story for link unfurl previews (Notion, docs, GitHub, etc.) which Slack puts in the 'attachments' array. Two fixes in the inbound handler: 1. _extract_text_from_slack_blocks walks rich_text / rich_text_quote / rich_text_list / rich_text_preformatted trees and renders readable text ('> quoted', '• bullet', code fences), dedupes against the plain text field, and appends the extracted content so the agent sees everything. 2. Link unfurl / attachment preview extraction reads title, url, body, and footer from the 'attachments' array and appends a '📎 [title](url)\n body\n _footer_' section per preview. Skips is_msg_unfurl to avoid echoing our own Slack replies back. Routing is careful not to trust augmented text: mention gating (is_mentioned) and slash-command detection both run against the original 'text' field, so forwarded content containing '<@bot>' or '/deploy' in a quote can't trick the bot into responding in a channel it shouldn't or classifying a normal message as a command. Adjustment from original PR: dropped _serialize_slack_blocks_for_agent, which inlined a redacted JSON dump of non-rich_text blocks (section, accessory, actions, etc.) — the agent would see the raw Block Kit structure for UI-heavy alerts. It added up to 6000 characters to the prompt context on every qualifying message with no opt-out. The rich_text extraction and attachment unfurls cover the common bug-fix case (quoted/forwarded content + link previews) without the prefill tax. If a user needs block inspection later, it can return as a config opt-in. Also updates the Slack platform notes in session.py to accurately describe what the gateway inlines.
2026-04-26 13:02:27 -07:00
"Do not promise to perform these actions. The gateway may inline the "
"current message's Slack block/attachment payload when available, but "
"you still cannot call Slack APIs yourself."
)
elif context.source.platform == Platform.DISCORD:
# Inject the Discord IDs block only when the agent actually has
# Discord tools loaded this session — i.e. the user opted into
# `discord` / `discord_admin` via `hermes tools` AND the bot
# token is configured. Otherwise keep the stale-API disclaimer
# honest so we never promise tools the agent lacks.
if _discord_tools_loaded():
src = context.source
id_lines = ["", "**Discord IDs (for the `discord` / `discord_admin` tools):**"]
if src.guild_id:
id_lines.append(f" - Guild: `{src.guild_id}`")
if src.thread_id and src.parent_chat_id:
id_lines.append(f" - Parent channel: `{src.parent_chat_id}`")
id_lines.append(f" - Thread: `{src.thread_id}` (use as `channel_id` for fetch_messages etc.)")
else:
id_lines.append(f" - Channel: `{src.chat_id}`")
if src.message_id:
id_lines.append(f" - Triggering message: `{src.message_id}`")
lines.extend(id_lines)
else:
lines.append("")
lines.append(
"**Platform notes:** You are running inside Discord. "
"You do NOT have access to Discord-specific APIs — you cannot search "
"channel history, pin messages, manage roles, or list server members. "
"Do not promise to perform these actions. If the user asks, explain "
"that you can only read messages sent directly to you and respond."
)
elif context.source.platform == Platform.BLUEBUBBLES:
lines.append("")
lines.append(
"**Platform notes:** You are responding via iMessage. "
"Keep responses short and conversational — think texts, not essays. "
"Structure longer replies as separate short thoughts, each separated "
"by a blank line (double newline). Each block between blank lines "
"will be delivered as its own iMessage bubble, so write accordingly: "
"one idea per bubble, 13 sentences each. "
"If the user needs a detailed answer, give the short version first "
"and offer to elaborate."
)
elif context.source.platform == Platform.YUANBAO:
lines.append("")
lines.append(
"**Platform notes:** You are running inside Yuanbao. "
"You CAN send private (DM) messages via the send_message tool. "
"Use target='yuanbao:direct:<account_id>' for DM "
"and target='yuanbao:group:<group_code>' for group chat."
)
# Connected platforms
platforms_list = ["local (files on this machine)"]
for p in context.connected_platforms:
if p != Platform.LOCAL:
platforms_list.append(f"{p.value}: Connected ✓")
lines.append(f"**Connected Platforms:** {', '.join(platforms_list)}")
# Home channels
if context.home_channels:
lines.append("")
lines.append("**Home Channels (default destinations):**")
for platform, home in context.home_channels.items():
hc_id = _hash_chat_id(home.chat_id) if redact_pii else home.chat_id
lines.append(f" - {platform.value}: {home.name} (ID: {hc_id})")
# Delivery options for scheduled tasks
lines.append("")
lines.append("**Delivery options for scheduled tasks:**")
from hermes_constants import display_hermes_home
# Origin delivery
if context.source.platform == Platform.LOCAL:
lines.append("- `\"origin\"` → Local output (saved to files)")
else:
_origin_label = context.source.chat_name or (
_hash_chat_id(context.source.chat_id) if redact_pii else context.source.chat_id
)
lines.append(f"- `\"origin\"` → Back to this chat ({_origin_label})")
# Local always available
lines.append(
f"- `\"local\"` → Save to local files only ({display_hermes_home()}/cron/output/)"
)
# Platform home channels
for platform, home in context.home_channels.items():
lines.append(f"- `\"{platform.value}\"` → Home channel ({home.name})")
# Note about explicit targeting
lines.append("")
lines.append("*For explicit targeting, use `\"platform:chat_id\"` format if the user provides a specific chat ID.*")
return "\n".join(lines)
@dataclass
class SessionEntry:
"""
Entry in the session store.
Maps a session key to its current session ID and metadata.
"""
session_key: str
session_id: str
created_at: datetime
updated_at: datetime
# Origin metadata for delivery routing
origin: Optional[SessionSource] = None
# Display metadata
display_name: Optional[str] = None
platform: Optional[Platform] = None
chat_type: str = "dm"
# Token tracking
input_tokens: int = 0
output_tokens: int = 0
cache_read_tokens: int = 0
cache_write_tokens: int = 0
total_tokens: int = 0
estimated_cost_usd: float = 0.0
cost_status: str = "unknown"
# Last API-reported prompt tokens (for accurate compression pre-check)
last_prompt_tokens: int = 0
2026-02-22 02:16:11 -08:00
# Set when a session was created because the previous one expired;
# consumed once by the message handler to inject a notice into context
was_auto_reset: bool = False
auto_reset_reason: Optional[str] = None # "idle" or "daily"
reset_had_activity: bool = False # whether the expired session had any messages
2026-02-22 02:16:11 -08:00
refactor(memory): remove flush_memories entirely (#15696) The AIAgent.flush_memories pre-compression save, the gateway _flush_memories_for_session, and everything feeding them are obsolete now that the background memory/skill review handles persistent memory extraction. Problems with flush_memories: - Pre-dates the background review loop. It was the only memory-save path when introduced; the background review now fires every 10 user turns on CLI and gateway alike, which is far more frequent than compression or session reset ever triggered flush. - Blocking and synchronous. Pre-compression flush ran on the live agent before compression, blocking the user-visible response. - Cache-breaking. Flush built a temporary conversation prefix (system prompt + memory-only tool list) that diverged from the live conversation's cached prefix, invalidating prompt caching. The gateway variant spawned a fresh AIAgent with its own clean prompt for each finalized session — still cache-breaking, just in a different process. - Redundant. Background review runs in the live conversation's session context, gets the same content, writes to the same memory store, and doesn't break the cache. Everything flush_memories claimed to preserve is already covered. What this removes: - AIAgent.flush_memories() method (~248 LOC in run_agent.py) - Pre-compression flush call in _compress_context - flush_memories call sites in cli.py (/new + exit) - GatewayRunner._flush_memories_for_session + _async_flush_memories (and the 3 call sites: session expiry watcher, /new, /resume) - 'flush_memories' entry from DEFAULT_CONFIG auxiliary tasks, hermes tools UI task list, auxiliary_client docstrings - _memory_flush_min_turns config + init - #15631's headroom-deduction math in _check_compression_model_feasibility (headroom was only needed because flush dragged the full main-agent system prompt along; the compression summariser sends a single user-role prompt so new_threshold = aux_context is safe again) - The dedicated test files and assertions that exercised flush-specific paths What this renames (with read-time backcompat on sessions.json): - SessionEntry.memory_flushed -> SessionEntry.expiry_finalized. The session-expiry watcher still uses the flag to avoid re-running finalize/eviction on the same expired session; the new name reflects what it now actually gates. from_dict() reads 'expiry_finalized' first, falls back to the legacy 'memory_flushed' key so existing sessions.json files upgrade seamlessly. Supersedes #15631 and #15638. Tested: 383 targeted tests pass across run_agent/, agent/, cli/, and gateway/ session-boundary suites. No behavior regressions — background memory review continues to handle persistent memory extraction on both CLI and gateway.
2026-04-25 08:21:14 -07:00
# Set by the background expiry watcher after it finalizes an expired
# session (invoking on_session_finalize hooks and evicting the cached
# agent). Persisted to sessions.json so the flag survives gateway
# restarts — prevents redundant finalization runs.
expiry_finalized: bool = False
# When True the next call to get_or_create_session() will auto-reset
# this session (create a new session_id) so the user starts fresh.
# Set by /stop to break stuck-resume loops (#7536).
suspended: bool = False
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
# When True the session was interrupted by a gateway restart/shutdown
# drain timeout, but recovery is still expected. Unlike ``suspended``,
# ``resume_pending`` preserves the existing session_id on next access —
# the user stays on the same transcript and the agent auto-continues
# from where it left off. Cleared after the next successful turn.
# Escalation to ``suspended`` is handled by the existing
# ``.restart_failure_counts`` stuck-loop counter (#7536), not by a
# parallel counter on this entry.
resume_pending: bool = False
resume_reason: Optional[str] = None # e.g. "restart_timeout"
last_resume_marked_at: Optional[datetime] = None
def to_dict(self) -> Dict[str, Any]:
result = {
"session_key": self.session_key,
"session_id": self.session_id,
"created_at": self.created_at.isoformat(),
"updated_at": self.updated_at.isoformat(),
"display_name": self.display_name,
"platform": self.platform.value if self.platform else None,
"chat_type": self.chat_type,
"input_tokens": self.input_tokens,
"output_tokens": self.output_tokens,
"cache_read_tokens": self.cache_read_tokens,
"cache_write_tokens": self.cache_write_tokens,
"total_tokens": self.total_tokens,
"last_prompt_tokens": self.last_prompt_tokens,
"estimated_cost_usd": self.estimated_cost_usd,
"cost_status": self.cost_status,
refactor(memory): remove flush_memories entirely (#15696) The AIAgent.flush_memories pre-compression save, the gateway _flush_memories_for_session, and everything feeding them are obsolete now that the background memory/skill review handles persistent memory extraction. Problems with flush_memories: - Pre-dates the background review loop. It was the only memory-save path when introduced; the background review now fires every 10 user turns on CLI and gateway alike, which is far more frequent than compression or session reset ever triggered flush. - Blocking and synchronous. Pre-compression flush ran on the live agent before compression, blocking the user-visible response. - Cache-breaking. Flush built a temporary conversation prefix (system prompt + memory-only tool list) that diverged from the live conversation's cached prefix, invalidating prompt caching. The gateway variant spawned a fresh AIAgent with its own clean prompt for each finalized session — still cache-breaking, just in a different process. - Redundant. Background review runs in the live conversation's session context, gets the same content, writes to the same memory store, and doesn't break the cache. Everything flush_memories claimed to preserve is already covered. What this removes: - AIAgent.flush_memories() method (~248 LOC in run_agent.py) - Pre-compression flush call in _compress_context - flush_memories call sites in cli.py (/new + exit) - GatewayRunner._flush_memories_for_session + _async_flush_memories (and the 3 call sites: session expiry watcher, /new, /resume) - 'flush_memories' entry from DEFAULT_CONFIG auxiliary tasks, hermes tools UI task list, auxiliary_client docstrings - _memory_flush_min_turns config + init - #15631's headroom-deduction math in _check_compression_model_feasibility (headroom was only needed because flush dragged the full main-agent system prompt along; the compression summariser sends a single user-role prompt so new_threshold = aux_context is safe again) - The dedicated test files and assertions that exercised flush-specific paths What this renames (with read-time backcompat on sessions.json): - SessionEntry.memory_flushed -> SessionEntry.expiry_finalized. The session-expiry watcher still uses the flag to avoid re-running finalize/eviction on the same expired session; the new name reflects what it now actually gates. from_dict() reads 'expiry_finalized' first, falls back to the legacy 'memory_flushed' key so existing sessions.json files upgrade seamlessly. Supersedes #15631 and #15638. Tested: 383 targeted tests pass across run_agent/, agent/, cli/, and gateway/ session-boundary suites. No behavior regressions — background memory review continues to handle persistent memory extraction on both CLI and gateway.
2026-04-25 08:21:14 -07:00
"expiry_finalized": self.expiry_finalized,
"suspended": self.suspended,
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
"resume_pending": self.resume_pending,
"resume_reason": self.resume_reason,
"last_resume_marked_at": (
self.last_resume_marked_at.isoformat()
if self.last_resume_marked_at
else None
),
}
if self.origin:
result["origin"] = self.origin.to_dict()
return result
@classmethod
def from_dict(cls, data: Dict[str, Any]) -> "SessionEntry":
origin = None
if "origin" in data and data["origin"]:
origin = SessionSource.from_dict(data["origin"])
platform = None
if data.get("platform"):
try:
platform = Platform(data["platform"])
except ValueError as e:
logger.debug("Unknown platform value %r: %s", data["platform"], e)
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
last_resume_marked_at = None
_lrma = data.get("last_resume_marked_at")
if _lrma:
try:
last_resume_marked_at = datetime.fromisoformat(_lrma)
except (TypeError, ValueError):
last_resume_marked_at = None
return cls(
session_key=data["session_key"],
session_id=data["session_id"],
created_at=datetime.fromisoformat(data["created_at"]),
updated_at=datetime.fromisoformat(data["updated_at"]),
origin=origin,
display_name=data.get("display_name"),
platform=platform,
chat_type=data.get("chat_type", "dm"),
input_tokens=data.get("input_tokens", 0),
output_tokens=data.get("output_tokens", 0),
cache_read_tokens=data.get("cache_read_tokens", 0),
cache_write_tokens=data.get("cache_write_tokens", 0),
total_tokens=data.get("total_tokens", 0),
last_prompt_tokens=data.get("last_prompt_tokens", 0),
estimated_cost_usd=data.get("estimated_cost_usd", 0.0),
cost_status=data.get("cost_status", "unknown"),
refactor(memory): remove flush_memories entirely (#15696) The AIAgent.flush_memories pre-compression save, the gateway _flush_memories_for_session, and everything feeding them are obsolete now that the background memory/skill review handles persistent memory extraction. Problems with flush_memories: - Pre-dates the background review loop. It was the only memory-save path when introduced; the background review now fires every 10 user turns on CLI and gateway alike, which is far more frequent than compression or session reset ever triggered flush. - Blocking and synchronous. Pre-compression flush ran on the live agent before compression, blocking the user-visible response. - Cache-breaking. Flush built a temporary conversation prefix (system prompt + memory-only tool list) that diverged from the live conversation's cached prefix, invalidating prompt caching. The gateway variant spawned a fresh AIAgent with its own clean prompt for each finalized session — still cache-breaking, just in a different process. - Redundant. Background review runs in the live conversation's session context, gets the same content, writes to the same memory store, and doesn't break the cache. Everything flush_memories claimed to preserve is already covered. What this removes: - AIAgent.flush_memories() method (~248 LOC in run_agent.py) - Pre-compression flush call in _compress_context - flush_memories call sites in cli.py (/new + exit) - GatewayRunner._flush_memories_for_session + _async_flush_memories (and the 3 call sites: session expiry watcher, /new, /resume) - 'flush_memories' entry from DEFAULT_CONFIG auxiliary tasks, hermes tools UI task list, auxiliary_client docstrings - _memory_flush_min_turns config + init - #15631's headroom-deduction math in _check_compression_model_feasibility (headroom was only needed because flush dragged the full main-agent system prompt along; the compression summariser sends a single user-role prompt so new_threshold = aux_context is safe again) - The dedicated test files and assertions that exercised flush-specific paths What this renames (with read-time backcompat on sessions.json): - SessionEntry.memory_flushed -> SessionEntry.expiry_finalized. The session-expiry watcher still uses the flag to avoid re-running finalize/eviction on the same expired session; the new name reflects what it now actually gates. from_dict() reads 'expiry_finalized' first, falls back to the legacy 'memory_flushed' key so existing sessions.json files upgrade seamlessly. Supersedes #15631 and #15638. Tested: 383 targeted tests pass across run_agent/, agent/, cli/, and gateway/ session-boundary suites. No behavior regressions — background memory review continues to handle persistent memory extraction on both CLI and gateway.
2026-04-25 08:21:14 -07:00
expiry_finalized=data.get("expiry_finalized", data.get("memory_flushed", False)),
suspended=data.get("suspended", False),
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
resume_pending=data.get("resume_pending", False),
resume_reason=data.get("resume_reason"),
last_resume_marked_at=last_resume_marked_at,
)
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
def is_shared_multi_user_session(
source: SessionSource,
*,
group_sessions_per_user: bool = True,
thread_sessions_per_user: bool = False,
) -> bool:
"""Return True when a non-DM session is shared across participants.
Mirrors the isolation rules in :func:`build_session_key`:
- DMs are never shared.
- Threads are shared unless ``thread_sessions_per_user`` is True.
- Non-thread group/channel sessions are shared unless
``group_sessions_per_user`` is True (default: True = isolated).
"""
if source.chat_type == "dm":
return False
if source.thread_id:
return not thread_sessions_per_user
return not group_sessions_per_user
def build_session_key(
source: SessionSource,
group_sessions_per_user: bool = True,
thread_sessions_per_user: bool = False,
) -> str:
"""Build a deterministic session key from a message source.
This is the single source of truth for session key construction.
DM rules:
- DMs include chat_id when present, so each private conversation is isolated.
- thread_id further differentiates threaded DMs within the same DM chat.
- Without chat_id, thread_id is used as a best-effort fallback.
- Without thread_id or chat_id, DMs share a single session.
Group/channel rules:
- chat_id identifies the parent group/channel.
- user_id/user_id_alt isolates participants within that parent chat when available when
``group_sessions_per_user`` is enabled.
- thread_id differentiates threads within that parent chat. When
``thread_sessions_per_user`` is False (default), threads are *shared* across all
participants user_id is NOT appended, so every user in the thread
shares a single session. This is the expected UX for threaded
conversations (Telegram forum topics, Discord threads, Slack threads).
- Without participant identifiers, or when isolation is disabled, messages fall back to one
shared session per chat.
- Without identifiers, messages fall back to one session per platform/chat_type.
"""
platform = source.platform.value
if source.chat_type == "dm":
fix(gateway): canonicalize WhatsApp identity in session keys Hermes' WhatsApp bridge routinely surfaces the same person under either a phone-format JID (60123456789@s.whatsapp.net) or a LID (…@lid), and may flip between the two for a single human within the same conversation. Before this change, build_session_key used the raw identifier verbatim, so the bridge reshuffling an alias form produced two distinct session keys for the same person — in two places: 1. DM chat_id — a user's DM sessions split in half, transcripts and per-sender state diverge. 2. Group participant_id (with group_sessions_per_user enabled) — a member's per-user session inside a group splits in half for the same reason. Add a canonicalizer that walks the bridge's lid-mapping-*.json files and picks the shortest/numeric-preferred alias as the stable identity. build_session_key now routes both the DM chat_id and the group participant_id through this helper when the platform is WhatsApp. All other platforms and chat types are untouched. Expose canonical_whatsapp_identifier and normalize_whatsapp_identifier as public helpers. Plugins that need per-sender behaviour (role-based routing, per-contact authorization, policy gating) need the same identity resolution Hermes uses internally; without a public helper, each plugin would have to re-implement the walker against the bridge's internal on-disk format. Keeping this alongside build_session_key makes it authoritative and one refactor away if the bridge ever changes shape. _expand_whatsapp_aliases stays private — it's an implementation detail of how the mapping files are walked, not a contract callers should depend on.
2026-04-24 18:40:20 +08:00
dm_chat_id = source.chat_id
if source.platform == Platform.WHATSAPP:
dm_chat_id = canonical_whatsapp_identifier(source.chat_id)
if dm_chat_id:
if source.thread_id:
fix(gateway): canonicalize WhatsApp identity in session keys Hermes' WhatsApp bridge routinely surfaces the same person under either a phone-format JID (60123456789@s.whatsapp.net) or a LID (…@lid), and may flip between the two for a single human within the same conversation. Before this change, build_session_key used the raw identifier verbatim, so the bridge reshuffling an alias form produced two distinct session keys for the same person — in two places: 1. DM chat_id — a user's DM sessions split in half, transcripts and per-sender state diverge. 2. Group participant_id (with group_sessions_per_user enabled) — a member's per-user session inside a group splits in half for the same reason. Add a canonicalizer that walks the bridge's lid-mapping-*.json files and picks the shortest/numeric-preferred alias as the stable identity. build_session_key now routes both the DM chat_id and the group participant_id through this helper when the platform is WhatsApp. All other platforms and chat types are untouched. Expose canonical_whatsapp_identifier and normalize_whatsapp_identifier as public helpers. Plugins that need per-sender behaviour (role-based routing, per-contact authorization, policy gating) need the same identity resolution Hermes uses internally; without a public helper, each plugin would have to re-implement the walker against the bridge's internal on-disk format. Keeping this alongside build_session_key makes it authoritative and one refactor away if the bridge ever changes shape. _expand_whatsapp_aliases stays private — it's an implementation detail of how the mapping files are walked, not a contract callers should depend on.
2026-04-24 18:40:20 +08:00
return f"agent:main:{platform}:dm:{dm_chat_id}:{source.thread_id}"
return f"agent:main:{platform}:dm:{dm_chat_id}"
if source.thread_id:
return f"agent:main:{platform}:dm:{source.thread_id}"
return f"agent:main:{platform}:dm"
participant_id = source.user_id_alt or source.user_id
fix(gateway): canonicalize WhatsApp identity in session keys Hermes' WhatsApp bridge routinely surfaces the same person under either a phone-format JID (60123456789@s.whatsapp.net) or a LID (…@lid), and may flip between the two for a single human within the same conversation. Before this change, build_session_key used the raw identifier verbatim, so the bridge reshuffling an alias form produced two distinct session keys for the same person — in two places: 1. DM chat_id — a user's DM sessions split in half, transcripts and per-sender state diverge. 2. Group participant_id (with group_sessions_per_user enabled) — a member's per-user session inside a group splits in half for the same reason. Add a canonicalizer that walks the bridge's lid-mapping-*.json files and picks the shortest/numeric-preferred alias as the stable identity. build_session_key now routes both the DM chat_id and the group participant_id through this helper when the platform is WhatsApp. All other platforms and chat types are untouched. Expose canonical_whatsapp_identifier and normalize_whatsapp_identifier as public helpers. Plugins that need per-sender behaviour (role-based routing, per-contact authorization, policy gating) need the same identity resolution Hermes uses internally; without a public helper, each plugin would have to re-implement the walker against the bridge's internal on-disk format. Keeping this alongside build_session_key makes it authoritative and one refactor away if the bridge ever changes shape. _expand_whatsapp_aliases stays private — it's an implementation detail of how the mapping files are walked, not a contract callers should depend on.
2026-04-24 18:40:20 +08:00
if participant_id and source.platform == Platform.WHATSAPP:
# Same JID/LID-flip bug as the DM case: without canonicalisation, a
# single group member gets two isolated per-user sessions when the
# bridge reshuffles alias forms.
participant_id = canonical_whatsapp_identifier(str(participant_id)) or participant_id
key_parts = ["agent:main", platform, source.chat_type]
if source.chat_id:
key_parts.append(source.chat_id)
if source.thread_id:
key_parts.append(source.thread_id)
# In threads, default to shared sessions (all participants see the same
# conversation). Per-user isolation only applies when explicitly enabled
# via thread_sessions_per_user, or when there is no thread (regular group).
isolate_user = group_sessions_per_user
if source.thread_id and not thread_sessions_per_user:
isolate_user = False
if isolate_user and participant_id:
key_parts.append(str(participant_id))
return ":".join(key_parts)
class SessionStore:
"""
Manages session storage and retrieval.
Uses SQLite (via SessionDB) for session metadata and message transcripts.
Falls back to legacy JSONL files if SQLite is unavailable.
"""
Add background process management with process tool, wait, PTY, and stdin support New process registry and tool for managing long-running background processes across all terminal backends (local, Docker, Singularity, Modal, SSH). Process Registry (tools/process_registry.py): - ProcessSession tracking with rolling 200KB output buffer - spawn_local() with optional PTY via ptyprocess for interactive CLIs - spawn_via_env() for non-local backends (runs inside sandbox, never on host) - Background reader threads per process (Popen stdout or PTY) - wait() with timeout clamping, interrupt support, and transparent limit reporting - JSON checkpoint to ~/.hermes/processes.json for gateway crash recovery - Module-level singleton shared across agent loop, gateway, and RL Process Tool (model_tools.py): - 7 actions: list, poll, log, wait, kill, write, submit - Paired with terminal in all toolsets (CLI, messaging, RL) - Timeout clamping with transparent notes in response Terminal Tool Updates (tools/terminal_tool.py): - Replaced nohup background mode with registry spawn (returns session_id) - Added workdir parameter for per-command working directory - Added check_interval parameter for gateway auto-check watchers - Added pty parameter for interactive CLI tools (Codex, Claude Code) - Updated TERMINAL_TOOL_DESCRIPTION with full background workflow docs - Cleanup thread now respects active background processes (won't reap sandbox) Gateway Integration (gateway/run.py, session.py, config.py): - Session reset protection: sessions with active processes exempt from reset - Default idle timeout increased from 2 hours to 24 hours - from_dict fallback aligned to match (was 120, now 1440) - session_key env var propagated to process registry for session mapping - Crash recovery on gateway startup via checkpoint probe - check_interval watcher: asyncio task polls process, delivers updates to platform RL Safety (environments/): - tool_context.py cleanup() kills background processes on episode end - hermes_base_env.py warns when enabled_toolsets is None (loads all tools) - Process tool safe in RL via wait() blocking the agent loop Also: - Added ptyprocess as optional dependency (in pyproject.toml [pty] extra + [all]) - Fixed pre-existing bug: rl_test_inference missing from TOOL_TO_TOOLSET_MAP - Updated AGENTS.md with process management docs and project structure - Updated README.md terminal section with process management overview
2026-02-17 02:51:31 -08:00
def __init__(self, sessions_dir: Path, config: GatewayConfig,
has_active_processes_fn=None):
self.sessions_dir = sessions_dir
self.config = config
self._entries: Dict[str, SessionEntry] = {}
self._loaded = False
self._lock = threading.Lock()
Add background process management with process tool, wait, PTY, and stdin support New process registry and tool for managing long-running background processes across all terminal backends (local, Docker, Singularity, Modal, SSH). Process Registry (tools/process_registry.py): - ProcessSession tracking with rolling 200KB output buffer - spawn_local() with optional PTY via ptyprocess for interactive CLIs - spawn_via_env() for non-local backends (runs inside sandbox, never on host) - Background reader threads per process (Popen stdout or PTY) - wait() with timeout clamping, interrupt support, and transparent limit reporting - JSON checkpoint to ~/.hermes/processes.json for gateway crash recovery - Module-level singleton shared across agent loop, gateway, and RL Process Tool (model_tools.py): - 7 actions: list, poll, log, wait, kill, write, submit - Paired with terminal in all toolsets (CLI, messaging, RL) - Timeout clamping with transparent notes in response Terminal Tool Updates (tools/terminal_tool.py): - Replaced nohup background mode with registry spawn (returns session_id) - Added workdir parameter for per-command working directory - Added check_interval parameter for gateway auto-check watchers - Added pty parameter for interactive CLI tools (Codex, Claude Code) - Updated TERMINAL_TOOL_DESCRIPTION with full background workflow docs - Cleanup thread now respects active background processes (won't reap sandbox) Gateway Integration (gateway/run.py, session.py, config.py): - Session reset protection: sessions with active processes exempt from reset - Default idle timeout increased from 2 hours to 24 hours - from_dict fallback aligned to match (was 120, now 1440) - session_key env var propagated to process registry for session mapping - Crash recovery on gateway startup via checkpoint probe - check_interval watcher: asyncio task polls process, delivers updates to platform RL Safety (environments/): - tool_context.py cleanup() kills background processes on episode end - hermes_base_env.py warns when enabled_toolsets is None (loads all tools) - Process tool safe in RL via wait() blocking the agent loop Also: - Added ptyprocess as optional dependency (in pyproject.toml [pty] extra + [all]) - Fixed pre-existing bug: rl_test_inference missing from TOOL_TO_TOOLSET_MAP - Updated AGENTS.md with process management docs and project structure - Updated README.md terminal section with process management overview
2026-02-17 02:51:31 -08:00
self._has_active_processes_fn = has_active_processes_fn
# Initialize SQLite session database
self._db = None
try:
from hermes_state import SessionDB
self._db = SessionDB()
except Exception as e:
print(f"[gateway] Warning: SQLite session store unavailable, falling back to JSONL: {e}")
def _ensure_loaded(self) -> None:
"""Load sessions index from disk if not already loaded."""
with self._lock:
self._ensure_loaded_locked()
def _ensure_loaded_locked(self) -> None:
"""Load sessions index from disk. Must be called with self._lock held."""
if self._loaded:
return
self.sessions_dir.mkdir(parents=True, exist_ok=True)
sessions_file = self.sessions_dir / "sessions.json"
if sessions_file.exists():
try:
with open(sessions_file, "r", encoding="utf-8") as f:
data = json.load(f)
for key, entry_data in data.items():
try:
self._entries[key] = SessionEntry.from_dict(entry_data)
except (ValueError, KeyError):
# Skip entries with unknown/removed platform values
continue
except Exception as e:
print(f"[gateway] Warning: Failed to load sessions: {e}")
self._loaded = True
def _save(self) -> None:
"""Save sessions index to disk (kept for session key -> ID mapping)."""
import tempfile
self.sessions_dir.mkdir(parents=True, exist_ok=True)
sessions_file = self.sessions_dir / "sessions.json"
data = {key: entry.to_dict() for key, entry in self._entries.items()}
fd, tmp_path = tempfile.mkstemp(
dir=str(self.sessions_dir), suffix=".tmp", prefix=".sessions_"
)
try:
with os.fdopen(fd, "w", encoding="utf-8") as f:
json.dump(data, f, indent=2)
f.flush()
os.fsync(f.fileno())
os.replace(tmp_path, sessions_file)
except BaseException:
try:
os.unlink(tmp_path)
except OSError as e:
logger.debug("Could not remove temp file %s: %s", tmp_path, e)
raise
def _generate_session_key(self, source: SessionSource) -> str:
"""Generate a session key from a source."""
return build_session_key(
source,
group_sessions_per_user=getattr(self.config, "group_sessions_per_user", True),
thread_sessions_per_user=getattr(self.config, "thread_sessions_per_user", False),
)
def _is_session_expired(self, entry: SessionEntry) -> bool:
"""Check if a session has expired based on its reset policy.
Works from the entry alone no SessionSource needed.
Used by the background expiry watcher to proactively flush memories.
Sessions with active background processes are never considered expired.
"""
if self._has_active_processes_fn:
if self._has_active_processes_fn(entry.session_key):
return False
policy = self.config.get_reset_policy(
platform=entry.platform,
session_type=entry.chat_type,
)
if policy.mode == "none":
return False
now = _now()
if policy.mode in ("idle", "both"):
idle_deadline = entry.updated_at + timedelta(minutes=policy.idle_minutes)
if now > idle_deadline:
return True
if policy.mode in ("daily", "both"):
today_reset = now.replace(
hour=policy.at_hour,
minute=0, second=0, microsecond=0,
)
if now.hour < policy.at_hour:
today_reset -= timedelta(days=1)
if entry.updated_at < today_reset:
return True
return False
def _should_reset(self, entry: SessionEntry, source: SessionSource) -> Optional[str]:
"""
Check if a session should be reset based on policy.
Returns the reset reason ("idle" or "daily") if a reset is needed,
or None if the session is still valid.
Add background process management with process tool, wait, PTY, and stdin support New process registry and tool for managing long-running background processes across all terminal backends (local, Docker, Singularity, Modal, SSH). Process Registry (tools/process_registry.py): - ProcessSession tracking with rolling 200KB output buffer - spawn_local() with optional PTY via ptyprocess for interactive CLIs - spawn_via_env() for non-local backends (runs inside sandbox, never on host) - Background reader threads per process (Popen stdout or PTY) - wait() with timeout clamping, interrupt support, and transparent limit reporting - JSON checkpoint to ~/.hermes/processes.json for gateway crash recovery - Module-level singleton shared across agent loop, gateway, and RL Process Tool (model_tools.py): - 7 actions: list, poll, log, wait, kill, write, submit - Paired with terminal in all toolsets (CLI, messaging, RL) - Timeout clamping with transparent notes in response Terminal Tool Updates (tools/terminal_tool.py): - Replaced nohup background mode with registry spawn (returns session_id) - Added workdir parameter for per-command working directory - Added check_interval parameter for gateway auto-check watchers - Added pty parameter for interactive CLI tools (Codex, Claude Code) - Updated TERMINAL_TOOL_DESCRIPTION with full background workflow docs - Cleanup thread now respects active background processes (won't reap sandbox) Gateway Integration (gateway/run.py, session.py, config.py): - Session reset protection: sessions with active processes exempt from reset - Default idle timeout increased from 2 hours to 24 hours - from_dict fallback aligned to match (was 120, now 1440) - session_key env var propagated to process registry for session mapping - Crash recovery on gateway startup via checkpoint probe - check_interval watcher: asyncio task polls process, delivers updates to platform RL Safety (environments/): - tool_context.py cleanup() kills background processes on episode end - hermes_base_env.py warns when enabled_toolsets is None (loads all tools) - Process tool safe in RL via wait() blocking the agent loop Also: - Added ptyprocess as optional dependency (in pyproject.toml [pty] extra + [all]) - Fixed pre-existing bug: rl_test_inference missing from TOOL_TO_TOOLSET_MAP - Updated AGENTS.md with process management docs and project structure - Updated README.md terminal section with process management overview
2026-02-17 02:51:31 -08:00
Sessions with active background processes are never reset.
"""
Add background process management with process tool, wait, PTY, and stdin support New process registry and tool for managing long-running background processes across all terminal backends (local, Docker, Singularity, Modal, SSH). Process Registry (tools/process_registry.py): - ProcessSession tracking with rolling 200KB output buffer - spawn_local() with optional PTY via ptyprocess for interactive CLIs - spawn_via_env() for non-local backends (runs inside sandbox, never on host) - Background reader threads per process (Popen stdout or PTY) - wait() with timeout clamping, interrupt support, and transparent limit reporting - JSON checkpoint to ~/.hermes/processes.json for gateway crash recovery - Module-level singleton shared across agent loop, gateway, and RL Process Tool (model_tools.py): - 7 actions: list, poll, log, wait, kill, write, submit - Paired with terminal in all toolsets (CLI, messaging, RL) - Timeout clamping with transparent notes in response Terminal Tool Updates (tools/terminal_tool.py): - Replaced nohup background mode with registry spawn (returns session_id) - Added workdir parameter for per-command working directory - Added check_interval parameter for gateway auto-check watchers - Added pty parameter for interactive CLI tools (Codex, Claude Code) - Updated TERMINAL_TOOL_DESCRIPTION with full background workflow docs - Cleanup thread now respects active background processes (won't reap sandbox) Gateway Integration (gateway/run.py, session.py, config.py): - Session reset protection: sessions with active processes exempt from reset - Default idle timeout increased from 2 hours to 24 hours - from_dict fallback aligned to match (was 120, now 1440) - session_key env var propagated to process registry for session mapping - Crash recovery on gateway startup via checkpoint probe - check_interval watcher: asyncio task polls process, delivers updates to platform RL Safety (environments/): - tool_context.py cleanup() kills background processes on episode end - hermes_base_env.py warns when enabled_toolsets is None (loads all tools) - Process tool safe in RL via wait() blocking the agent loop Also: - Added ptyprocess as optional dependency (in pyproject.toml [pty] extra + [all]) - Fixed pre-existing bug: rl_test_inference missing from TOOL_TO_TOOLSET_MAP - Updated AGENTS.md with process management docs and project structure - Updated README.md terminal section with process management overview
2026-02-17 02:51:31 -08:00
if self._has_active_processes_fn:
session_key = self._generate_session_key(source)
if self._has_active_processes_fn(session_key):
return None
Add background process management with process tool, wait, PTY, and stdin support New process registry and tool for managing long-running background processes across all terminal backends (local, Docker, Singularity, Modal, SSH). Process Registry (tools/process_registry.py): - ProcessSession tracking with rolling 200KB output buffer - spawn_local() with optional PTY via ptyprocess for interactive CLIs - spawn_via_env() for non-local backends (runs inside sandbox, never on host) - Background reader threads per process (Popen stdout or PTY) - wait() with timeout clamping, interrupt support, and transparent limit reporting - JSON checkpoint to ~/.hermes/processes.json for gateway crash recovery - Module-level singleton shared across agent loop, gateway, and RL Process Tool (model_tools.py): - 7 actions: list, poll, log, wait, kill, write, submit - Paired with terminal in all toolsets (CLI, messaging, RL) - Timeout clamping with transparent notes in response Terminal Tool Updates (tools/terminal_tool.py): - Replaced nohup background mode with registry spawn (returns session_id) - Added workdir parameter for per-command working directory - Added check_interval parameter for gateway auto-check watchers - Added pty parameter for interactive CLI tools (Codex, Claude Code) - Updated TERMINAL_TOOL_DESCRIPTION with full background workflow docs - Cleanup thread now respects active background processes (won't reap sandbox) Gateway Integration (gateway/run.py, session.py, config.py): - Session reset protection: sessions with active processes exempt from reset - Default idle timeout increased from 2 hours to 24 hours - from_dict fallback aligned to match (was 120, now 1440) - session_key env var propagated to process registry for session mapping - Crash recovery on gateway startup via checkpoint probe - check_interval watcher: asyncio task polls process, delivers updates to platform RL Safety (environments/): - tool_context.py cleanup() kills background processes on episode end - hermes_base_env.py warns when enabled_toolsets is None (loads all tools) - Process tool safe in RL via wait() blocking the agent loop Also: - Added ptyprocess as optional dependency (in pyproject.toml [pty] extra + [all]) - Fixed pre-existing bug: rl_test_inference missing from TOOL_TO_TOOLSET_MAP - Updated AGENTS.md with process management docs and project structure - Updated README.md terminal section with process management overview
2026-02-17 02:51:31 -08:00
policy = self.config.get_reset_policy(
platform=source.platform,
session_type=source.chat_type
)
if policy.mode == "none":
return None
now = _now()
if policy.mode in ("idle", "both"):
idle_deadline = entry.updated_at + timedelta(minutes=policy.idle_minutes)
if now > idle_deadline:
return "idle"
if policy.mode in ("daily", "both"):
today_reset = now.replace(
hour=policy.at_hour,
minute=0,
second=0,
microsecond=0
)
if now.hour < policy.at_hour:
today_reset -= timedelta(days=1)
if entry.updated_at < today_reset:
return "daily"
return None
def has_any_sessions(self) -> bool:
"""Check if any sessions have ever been created (across all platforms).
Uses the SQLite database as the source of truth because it preserves
historical session records (ended sessions still count). The in-memory
``_entries`` dict replaces entries on reset, so ``len(_entries)`` would
stay at 1 for single-platform users which is the bug this fixes.
The current session is already in the DB by the time this is called
(get_or_create_session runs first), so we check ``> 1``.
"""
if self._db:
try:
return self._db.session_count() > 1
except Exception:
pass # fall through to heuristic
# Fallback: check if sessions.json was loaded with existing data.
# This covers the rare case where the DB is unavailable.
with self._lock:
self._ensure_loaded_locked()
return len(self._entries) > 1
def get_or_create_session(
self,
source: SessionSource,
force_new: bool = False
) -> SessionEntry:
"""
Get an existing session or create a new one.
Evaluates reset policy to determine if the existing session is stale.
Creates a session record in SQLite when a new session starts.
"""
session_key = self._generate_session_key(source)
now = _now()
# SQLite calls are made outside the lock to avoid holding it during I/O.
# All _entries / _loaded mutations are protected by self._lock.
db_end_session_id = None
db_create_kwargs = None
with self._lock:
self._ensure_loaded_locked()
if session_key in self._entries and not force_new:
entry = self._entries[session_key]
# Auto-reset sessions marked as suspended (e.g. after /stop
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
# broke a stuck loop — #7536). ``suspended`` is the hard
# forced-wipe signal and always wins over ``resume_pending``,
# so repeated interrupted restarts that escalate via the
# existing ``.restart_failure_counts`` stuck-loop counter
# still converge to a clean slate.
if entry.suspended:
reset_reason = "suspended"
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
elif entry.resume_pending:
# Restart-interrupted session: preserve the session_id
# and return the existing entry so the transcript
# reloads intact. ``resume_pending`` is cleared after
# the NEXT successful turn completes (not here), which
# means a re-interrupted retry keeps trying — the
# stuck-loop counter handles terminal escalation.
entry.updated_at = now
self._save()
return entry
else:
reset_reason = self._should_reset(entry, source)
if not reset_reason:
entry.updated_at = now
self._save()
return entry
else:
fix(gateway): persist memory flush state to prevent redundant re-flushes on restart (#4481) * fix: force-close TCP sockets on client cleanup, detect and recover dead connections When a provider drops connections mid-stream (e.g. OpenRouter outage), httpx's graceful close leaves sockets in CLOSE-WAIT indefinitely. These zombie connections accumulate and can prevent recovery without restarting. Changes: - _force_close_tcp_sockets: walks the httpx connection pool and issues socket.shutdown(SHUT_RDWR) + close() to force TCP RST on every socket when a client is closed, preventing CLOSE-WAIT accumulation - _cleanup_dead_connections: probes the primary client's pool for dead sockets (recv MSG_PEEK), rebuilds the client if any are found - Pre-turn health check at the start of each run_conversation call that auto-recovers with a user-facing status message - Primary client rebuild after stale stream detection to purge pool - User-facing messages on streaming connection failures: "Connection to provider dropped — Reconnecting (attempt 2/3)" "Connection failed after 3 attempts — try again in a moment" Made-with: Cursor * fix: pool entry missing base_url for openrouter, clean error messages - _resolve_runtime_from_pool_entry: add OPENROUTER_BASE_URL fallback when pool entry has no runtime_base_url (pool entries from auth.json credential_pool often omit base_url) - Replace Rich console.print for auth errors with plain print() to prevent ANSI escape code mangling through prompt_toolkit's stdout patch - Force-close TCP sockets on client cleanup to prevent CLOSE-WAIT accumulation after provider outages - Pre-turn dead connection detection with auto-recovery and user message - Primary client rebuild after stale stream detection - User-facing status messages on streaming connection failures/retries Made-with: Cursor * fix(gateway): persist memory flush state to prevent redundant re-flushes on restart The _session_expiry_watcher tracked flushed sessions in an in-memory set (_pre_flushed_sessions) that was lost on gateway restart. Expired sessions remained in sessions.json and were re-discovered every restart, causing redundant AIAgent runs that burned API credits and blocked the event loop. Fix: Add a memory_flushed boolean field to SessionEntry, persisted in sessions.json. The watcher sets it after a successful flush. On restart, the flag survives and the watcher skips already-flushed sessions. - Add memory_flushed field to SessionEntry with to_dict/from_dict support - Old sessions.json entries without the field default to False (backward compat) - Remove the ephemeral _pre_flushed_sessions set from SessionStore - Update tests: save/load roundtrip, legacy entry compat, auto-reset behavior
2026-04-01 12:05:02 -07:00
# Session is being auto-reset.
was_auto_reset = True
auto_reset_reason = reset_reason
# Track whether the expired session had any real conversation
reset_had_activity = entry.total_tokens > 0
db_end_session_id = entry.session_id
else:
was_auto_reset = False
auto_reset_reason = None
reset_had_activity = False
# Create new session
session_id = f"{now.strftime('%Y%m%d_%H%M%S')}_{uuid.uuid4().hex[:8]}"
entry = SessionEntry(
session_key=session_key,
session_id=session_id,
created_at=now,
updated_at=now,
origin=source,
display_name=source.chat_name,
platform=source.platform,
chat_type=source.chat_type,
was_auto_reset=was_auto_reset,
auto_reset_reason=auto_reset_reason,
reset_had_activity=reset_had_activity,
)
self._entries[session_key] = entry
self._save()
db_create_kwargs = {
"session_id": session_id,
"source": source.platform.value,
"user_id": source.user_id,
}
# SQLite operations outside the lock
if self._db and db_end_session_id:
try:
self._db.end_session(db_end_session_id, "session_reset")
except Exception as e:
logger.debug("Session DB operation failed: %s", e)
if self._db and db_create_kwargs:
try:
self._db.create_session(**db_create_kwargs)
except Exception as e:
print(f"[gateway] Warning: Failed to create SQLite session: {e}")
return entry
def update_session(
self,
session_key: str,
last_prompt_tokens: int = None,
) -> None:
"""Update lightweight session metadata after an interaction."""
with self._lock:
self._ensure_loaded_locked()
if session_key in self._entries:
entry = self._entries[session_key]
entry.updated_at = _now()
if last_prompt_tokens is not None:
entry.last_prompt_tokens = last_prompt_tokens
self._save()
def suspend_session(self, session_key: str) -> bool:
"""Mark a session as suspended so it auto-resets on next access.
Used by ``/stop`` to prevent stuck sessions from being resumed
after a gateway restart (#7536). Returns True if the session
existed and was marked.
"""
with self._lock:
self._ensure_loaded_locked()
if session_key in self._entries:
self._entries[session_key].suspended = True
self._save()
return True
return False
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
def mark_resume_pending(
self,
session_key: str,
reason: str = "restart_timeout",
) -> bool:
"""Mark a session as resumable after a restart interruption.
Unlike ``suspend_session()``, this preserves the existing
``session_id`` and the transcript. The next call to
``get_or_create_session()`` for this key returns the same entry
so the user auto-resumes on the same conversation lane.
Returns True if the session existed and was marked.
"""
with self._lock:
self._ensure_loaded_locked()
if session_key in self._entries:
entry = self._entries[session_key]
# Never override an explicit ``suspended`` — that is a hard
# forced-wipe signal (from /stop or stuck-loop escalation).
if entry.suspended:
return False
entry.resume_pending = True
entry.resume_reason = reason
entry.last_resume_marked_at = _now()
self._save()
return True
return False
def clear_resume_pending(self, session_key: str) -> bool:
"""Clear the resume-pending flag after a successful resumed turn.
Called from the gateway after ``run_conversation()`` returns a
final response for a session that had ``resume_pending=True``,
signalling that recovery succeeded.
Returns True if a flag was cleared.
"""
with self._lock:
self._ensure_loaded_locked()
entry = self._entries.get(session_key)
if entry is None or not entry.resume_pending:
return False
entry.resume_pending = False
entry.resume_reason = None
entry.last_resume_marked_at = None
self._save()
return True
fix(gateway): prune stale SessionStore entries to bound memory + disk (#11789) SessionStore._entries grew unbounded. Every unique (platform, chat_id, thread_id, user_id) tuple ever seen was kept in RAM and rewritten to sessions.json on every message. A Discord bot in 100 servers x 100 channels x ~100 rotating users accumulates on the order of 10^5 entries after a few months; each sessions.json write becomes an O(n) fsync. Nothing trimmed this — there was no TTL, no cap, no eviction path. Changes ------- * SessionStore.prune_old_entries(max_age_days) — drops entries whose updated_at is older than the cutoff. Preserves: - suspended entries (user paused them via /stop for later resume) - entries with an active background process attached Pruning is functionally identical to a natural reset-policy expiry: SQLite transcript stays, session_key -> session_id mapping dropped, returning user gets a fresh session. * GatewayConfig.session_store_max_age_days (default 90; 0 disables). Serialized in to_dict/from_dict, coerced from bad types / negatives to safe defaults. No migration needed — missing field -> 90 days. * _session_expiry_watcher calls prune_old_entries once per hour (first tick is immediate). Uses the existing watcher loop so no new background task is created. Why not more aggressive ----------------------- 90 days is long enough that legitimate long-idle users (seasonal, vacation, etc.) aren't surprised — pruning just means they get a fresh session on return, same outcome they'd get from any other reset-policy trigger. Admins can lower it via config; 0 disables. Tests ----- tests/gateway/test_session_store_prune.py — 17 cases covering: * entry age based on updated_at, not created_at * max_age_days=0 disables; negative coerces to 0 * suspended + active-process entries are skipped * _save fires iff something was removed * disk JSON reflects post-prune state * thread safety against concurrent readers * config field roundtrips + graceful fallback on bad values * watcher gate logic (first tick prunes, subsequent within 1h don't) 119 broader session/gateway tests remain green.
2026-04-17 13:48:49 -07:00
def prune_old_entries(self, max_age_days: int) -> int:
"""Drop SessionEntry records older than max_age_days.
Pruning is based on ``updated_at`` (last activity), not ``created_at``.
A session that's been active within the window is kept regardless of
how old it is. Entries marked ``suspended`` are kept the user
explicitly paused them for later resume. Entries held by an active
process (via has_active_processes_fn) are also kept so long-running
background work isn't orphaned.
Pruning is functionally identical to a natural reset-policy expiry:
the transcript in SQLite stays, but the session_key session_id
mapping is dropped and the user starts a fresh session on return.
``max_age_days <= 0`` disables pruning; returns 0 immediately.
Returns the number of entries removed.
"""
if max_age_days is None or max_age_days <= 0:
return 0
from datetime import timedelta
cutoff = _now() - timedelta(days=max_age_days)
removed_keys: list[str] = []
with self._lock:
self._ensure_loaded_locked()
for key, entry in list(self._entries.items()):
if entry.suspended:
continue
# Never prune sessions with an active background process
# attached — the user may still be waiting on output.
# The callback is keyed by session_key (see process_registry.
# has_active_for_session); passing session_id here used to
# never match, so active sessions got pruned anyway.
fix(gateway): prune stale SessionStore entries to bound memory + disk (#11789) SessionStore._entries grew unbounded. Every unique (platform, chat_id, thread_id, user_id) tuple ever seen was kept in RAM and rewritten to sessions.json on every message. A Discord bot in 100 servers x 100 channels x ~100 rotating users accumulates on the order of 10^5 entries after a few months; each sessions.json write becomes an O(n) fsync. Nothing trimmed this — there was no TTL, no cap, no eviction path. Changes ------- * SessionStore.prune_old_entries(max_age_days) — drops entries whose updated_at is older than the cutoff. Preserves: - suspended entries (user paused them via /stop for later resume) - entries with an active background process attached Pruning is functionally identical to a natural reset-policy expiry: SQLite transcript stays, session_key -> session_id mapping dropped, returning user gets a fresh session. * GatewayConfig.session_store_max_age_days (default 90; 0 disables). Serialized in to_dict/from_dict, coerced from bad types / negatives to safe defaults. No migration needed — missing field -> 90 days. * _session_expiry_watcher calls prune_old_entries once per hour (first tick is immediate). Uses the existing watcher loop so no new background task is created. Why not more aggressive ----------------------- 90 days is long enough that legitimate long-idle users (seasonal, vacation, etc.) aren't surprised — pruning just means they get a fresh session on return, same outcome they'd get from any other reset-policy trigger. Admins can lower it via config; 0 disables. Tests ----- tests/gateway/test_session_store_prune.py — 17 cases covering: * entry age based on updated_at, not created_at * max_age_days=0 disables; negative coerces to 0 * suspended + active-process entries are skipped * _save fires iff something was removed * disk JSON reflects post-prune state * thread safety against concurrent readers * config field roundtrips + graceful fallback on bad values * watcher gate logic (first tick prunes, subsequent within 1h don't) 119 broader session/gateway tests remain green.
2026-04-17 13:48:49 -07:00
if self._has_active_processes_fn is not None:
try:
if self._has_active_processes_fn(entry.session_key):
fix(gateway): prune stale SessionStore entries to bound memory + disk (#11789) SessionStore._entries grew unbounded. Every unique (platform, chat_id, thread_id, user_id) tuple ever seen was kept in RAM and rewritten to sessions.json on every message. A Discord bot in 100 servers x 100 channels x ~100 rotating users accumulates on the order of 10^5 entries after a few months; each sessions.json write becomes an O(n) fsync. Nothing trimmed this — there was no TTL, no cap, no eviction path. Changes ------- * SessionStore.prune_old_entries(max_age_days) — drops entries whose updated_at is older than the cutoff. Preserves: - suspended entries (user paused them via /stop for later resume) - entries with an active background process attached Pruning is functionally identical to a natural reset-policy expiry: SQLite transcript stays, session_key -> session_id mapping dropped, returning user gets a fresh session. * GatewayConfig.session_store_max_age_days (default 90; 0 disables). Serialized in to_dict/from_dict, coerced from bad types / negatives to safe defaults. No migration needed — missing field -> 90 days. * _session_expiry_watcher calls prune_old_entries once per hour (first tick is immediate). Uses the existing watcher loop so no new background task is created. Why not more aggressive ----------------------- 90 days is long enough that legitimate long-idle users (seasonal, vacation, etc.) aren't surprised — pruning just means they get a fresh session on return, same outcome they'd get from any other reset-policy trigger. Admins can lower it via config; 0 disables. Tests ----- tests/gateway/test_session_store_prune.py — 17 cases covering: * entry age based on updated_at, not created_at * max_age_days=0 disables; negative coerces to 0 * suspended + active-process entries are skipped * _save fires iff something was removed * disk JSON reflects post-prune state * thread safety against concurrent readers * config field roundtrips + graceful fallback on bad values * watcher gate logic (first tick prunes, subsequent within 1h don't) 119 broader session/gateway tests remain green.
2026-04-17 13:48:49 -07:00
continue
except Exception as exc:
logger.debug(
"has_active_processes_fn raised during prune for %s: %s",
entry.session_key, exc,
)
fix(gateway): prune stale SessionStore entries to bound memory + disk (#11789) SessionStore._entries grew unbounded. Every unique (platform, chat_id, thread_id, user_id) tuple ever seen was kept in RAM and rewritten to sessions.json on every message. A Discord bot in 100 servers x 100 channels x ~100 rotating users accumulates on the order of 10^5 entries after a few months; each sessions.json write becomes an O(n) fsync. Nothing trimmed this — there was no TTL, no cap, no eviction path. Changes ------- * SessionStore.prune_old_entries(max_age_days) — drops entries whose updated_at is older than the cutoff. Preserves: - suspended entries (user paused them via /stop for later resume) - entries with an active background process attached Pruning is functionally identical to a natural reset-policy expiry: SQLite transcript stays, session_key -> session_id mapping dropped, returning user gets a fresh session. * GatewayConfig.session_store_max_age_days (default 90; 0 disables). Serialized in to_dict/from_dict, coerced from bad types / negatives to safe defaults. No migration needed — missing field -> 90 days. * _session_expiry_watcher calls prune_old_entries once per hour (first tick is immediate). Uses the existing watcher loop so no new background task is created. Why not more aggressive ----------------------- 90 days is long enough that legitimate long-idle users (seasonal, vacation, etc.) aren't surprised — pruning just means they get a fresh session on return, same outcome they'd get from any other reset-policy trigger. Admins can lower it via config; 0 disables. Tests ----- tests/gateway/test_session_store_prune.py — 17 cases covering: * entry age based on updated_at, not created_at * max_age_days=0 disables; negative coerces to 0 * suspended + active-process entries are skipped * _save fires iff something was removed * disk JSON reflects post-prune state * thread safety against concurrent readers * config field roundtrips + graceful fallback on bad values * watcher gate logic (first tick prunes, subsequent within 1h don't) 119 broader session/gateway tests remain green.
2026-04-17 13:48:49 -07:00
if entry.updated_at < cutoff:
removed_keys.append(key)
for key in removed_keys:
self._entries.pop(key, None)
if removed_keys:
self._save()
if removed_keys:
logger.info(
"SessionStore pruned %d entries older than %d days",
len(removed_keys), max_age_days,
)
return len(removed_keys)
def suspend_recently_active(self, max_age_seconds: int = 120) -> int:
"""Mark recently-active sessions as suspended.
Called on gateway startup to prevent sessions that were likely
in-flight when the gateway last exited from being blindly resumed
(#7536). Only suspends sessions updated within *max_age_seconds*
to avoid resetting long-idle sessions that are harmless to resume.
Returns the number of sessions that were suspended.
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
Entries flagged ``resume_pending=True`` are skipped those were
marked intentionally by the drain-timeout path as recoverable.
Terminal escalation for genuinely stuck ``resume_pending`` sessions
is handled by the existing ``.restart_failure_counts`` stuck-loop
counter, which runs after this method on startup.
"""
from datetime import timedelta
cutoff = _now() - timedelta(seconds=max_age_seconds)
count = 0
with self._lock:
self._ensure_loaded_locked()
for entry in self._entries.values():
fix(gateway): auto-resume sessions after drain-timeout restart (#11852) (#12301) The shutdown banner promised "send any message after restart to resume where you left off" but the code did the opposite: a drain-timeout restart skipped the .clean_shutdown marker, which made the next startup call suspend_recently_active(), which marked the session suspended, which made get_or_create_session() spawn a fresh session_id with a 'Session automatically reset. Use /resume...' notice — contradicting the banner. Introduce a resume_pending state on SessionEntry that is distinct from suspended. Drain-timeout shutdown flags active sessions resume_pending instead of letting startup-wide suspension destroy them. The next message on the same session_key preserves the session_id, reloads the transcript, and the agent receives a reason-aware restart-resume system note that subsumes the existing tool-tail auto-continue note (PR #9934). Terminal escalation still flows through the existing .restart_failure_counts stuck-loop counter (PR #7536, threshold 3) — no parallel counter on SessionEntry. suspended still wins over resume_pending in get_or_create_session() so genuinely stuck sessions converge to a clean slate. Spec: PR #11852 (BrennerSpear). Implementation follows the spec with the approved correction (reuse .restart_failure_counts rather than adding a resume_attempts field). Changes: - gateway/session.py: SessionEntry.resume_pending/resume_reason/ last_resume_marked_at + to_dict/from_dict; SessionStore .mark_resume_pending()/clear_resume_pending(); get_or_create_session() returns existing entry when resume_pending (suspended still wins); suspend_recently_active() skips resume_pending entries. - gateway/run.py: _stop_impl() drain-timeout branch marks active sessions resume_pending before _interrupt_running_agents(); _run_agent() injects reason-aware restart-resume system note that subsumes the tool-tail case; successful-turn cleanup also clears resume_pending next to _clear_restart_failure_count(); _notify_active_sessions_of_shutdown() softens the restart banner to 'I'll try to resume where you left off' (honest about stuck-loop escalation). - tests/gateway/test_restart_resume_pending.py: 29 new tests covering SessionEntry roundtrip, mark/clear helpers, get_or_create_session precedence (suspended > resume_pending), suspend_recently_active skip, drain-timeout mark reason (restart vs shutdown), system-note injection decision tree (including tool-tail subsumption), banner wording, and stuck-loop escalation override.
2026-04-18 17:32:17 -07:00
if entry.resume_pending:
continue
if not entry.suspended and entry.updated_at >= cutoff:
entry.suspended = True
count += 1
if count:
self._save()
return count
def reset_session(self, session_key: str) -> Optional[SessionEntry]:
"""Force reset a session, creating a new session ID."""
db_end_session_id = None
db_create_kwargs = None
new_entry = None
with self._lock:
self._ensure_loaded_locked()
if session_key not in self._entries:
return None
old_entry = self._entries[session_key]
db_end_session_id = old_entry.session_id
now = _now()
session_id = f"{now.strftime('%Y%m%d_%H%M%S')}_{uuid.uuid4().hex[:8]}"
new_entry = SessionEntry(
session_key=session_key,
session_id=session_id,
created_at=now,
updated_at=now,
origin=old_entry.origin,
display_name=old_entry.display_name,
platform=old_entry.platform,
chat_type=old_entry.chat_type,
)
self._entries[session_key] = new_entry
self._save()
db_create_kwargs = {
"session_id": session_id,
"source": old_entry.platform.value if old_entry.platform else "unknown",
"user_id": old_entry.origin.user_id if old_entry.origin else None,
}
if self._db and db_end_session_id:
try:
self._db.end_session(db_end_session_id, "session_reset")
except Exception as e:
logger.debug("Session DB operation failed: %s", e)
if self._db and db_create_kwargs:
try:
self._db.create_session(**db_create_kwargs)
except Exception as e:
logger.debug("Session DB operation failed: %s", e)
return new_entry
def switch_session(self, session_key: str, target_session_id: str) -> Optional[SessionEntry]:
"""Switch a session key to point at an existing session ID.
Used by ``/resume`` to restore a previously-named session.
Ends the current session in SQLite (like reset), but instead of
generating a fresh session ID, re-uses ``target_session_id`` so the
old transcript is loaded on the next message. If the target session was
previously ended, re-open it so gateway resume semantics match the CLI.
"""
db_end_session_id = None
new_entry = None
with self._lock:
self._ensure_loaded_locked()
if session_key not in self._entries:
return None
old_entry = self._entries[session_key]
# Don't switch if already on that session
if old_entry.session_id == target_session_id:
return old_entry
db_end_session_id = old_entry.session_id
now = _now()
new_entry = SessionEntry(
session_key=session_key,
session_id=target_session_id,
created_at=now,
updated_at=now,
origin=old_entry.origin,
display_name=old_entry.display_name,
platform=old_entry.platform,
chat_type=old_entry.chat_type,
)
self._entries[session_key] = new_entry
self._save()
if self._db and db_end_session_id:
try:
self._db.end_session(db_end_session_id, "session_switch")
except Exception as e:
logger.debug("Session DB end_session failed: %s", e)
if self._db:
try:
self._db.reopen_session(target_session_id)
except Exception as e:
logger.debug("Session DB reopen_session failed: %s", e)
return new_entry
def list_sessions(self, active_minutes: Optional[int] = None) -> List[SessionEntry]:
"""List all sessions, optionally filtered by activity."""
with self._lock:
self._ensure_loaded_locked()
entries = list(self._entries.values())
if active_minutes is not None:
cutoff = _now() - timedelta(minutes=active_minutes)
entries = [e for e in entries if e.updated_at >= cutoff]
entries.sort(key=lambda e: e.updated_at, reverse=True)
return entries
def get_transcript_path(self, session_id: str) -> Path:
"""Get the path to a session's legacy transcript file."""
return self.sessions_dir / f"{session_id}.jsonl"
def append_to_transcript(self, session_id: str, message: Dict[str, Any], skip_db: bool = False) -> None:
"""Append a message to a session's transcript (SQLite + legacy JSONL).
Args:
skip_db: When True, only write to JSONL and skip the SQLite write.
Used when the agent already persisted messages to SQLite
via its own _flush_messages_to_session_db(), preventing
the duplicate-write bug (#860).
"""
# Write to SQLite (unless the agent already handled it)
if self._db and not skip_db:
try:
self._db.append_message(
session_id=session_id,
role=message.get("role", "unknown"),
content=message.get("content"),
tool_name=message.get("tool_name"),
tool_calls=message.get("tool_calls"),
tool_call_id=message.get("tool_call_id"),
reasoning=message.get("reasoning") if message.get("role") == "assistant" else None,
reasoning_content=message.get("reasoning_content") if message.get("role") == "assistant" else None,
reasoning_details=message.get("reasoning_details") if message.get("role") == "assistant" else None,
codex_reasoning_items=message.get("codex_reasoning_items") if message.get("role") == "assistant" else None,
codex_message_items=message.get("codex_message_items") if message.get("role") == "assistant" else None,
)
except Exception as e:
logger.debug("Session DB operation failed: %s", e)
# Also write legacy JSONL (keeps existing tooling working during transition)
transcript_path = self.get_transcript_path(session_id)
with open(transcript_path, "a", encoding="utf-8") as f:
f.write(json.dumps(message, ensure_ascii=False) + "\n")
def rewrite_transcript(self, session_id: str, messages: List[Dict[str, Any]]) -> None:
"""Replace the entire transcript for a session with new messages.
Used by /retry, /undo, and /compress to persist modified conversation history.
Rewrites both SQLite and legacy JSONL storage.
"""
# SQLite: clear old messages and re-insert
if self._db:
try:
self._db.clear_messages(session_id)
for msg in messages:
role = msg.get("role", "unknown")
self._db.append_message(
session_id=session_id,
role=role,
content=msg.get("content"),
tool_name=msg.get("tool_name"),
tool_calls=msg.get("tool_calls"),
tool_call_id=msg.get("tool_call_id"),
reasoning=msg.get("reasoning") if role == "assistant" else None,
reasoning_content=msg.get("reasoning_content") if role == "assistant" else None,
reasoning_details=msg.get("reasoning_details") if role == "assistant" else None,
codex_reasoning_items=msg.get("codex_reasoning_items") if role == "assistant" else None,
codex_message_items=msg.get("codex_message_items") if role == "assistant" else None,
)
except Exception as e:
logger.debug("Failed to rewrite transcript in DB: %s", e)
# JSONL: overwrite the file
transcript_path = self.get_transcript_path(session_id)
with open(transcript_path, "w", encoding="utf-8") as f:
for msg in messages:
f.write(json.dumps(msg, ensure_ascii=False) + "\n")
def load_transcript(self, session_id: str) -> List[Dict[str, Any]]:
"""Load all messages from a session's transcript."""
fix(state): SQLite concurrency hardening + session transcript integrity (#3249) * fix(session-db): survive CLI/gateway concurrent write contention Closes #3139 Three layered fixes for the scenario where CLI and gateway write to state.db concurrently, causing create_session() to fail with 'database is locked' and permanently disabling session_search on the gateway side. 1. Increase SQLite connection timeout: 10s -> 30s hermes_state.py: longer window for the WAL writer to finish a batch flush before the other process gives up entirely. 2. INSERT OR IGNORE in create_session hermes_state.py: prevents IntegrityError on duplicate session IDs (e.g. gateway restarts while CLI session is still alive). 3. Don't null out _session_db on create_session failure (main fix) run_agent.py: a transient lock at agent startup must not permanently disable session_search for the lifetime of that agent instance. _session_db now stays alive so subsequent flushes and searches work once the lock clears. 4. New ensure_session() helper + call it during flush hermes_state.py: INSERT OR IGNORE for a minimal session row. run_agent.py _flush_messages_to_session_db: calls ensure_session() before appending messages, so the FK constraint is satisfied even when create_session() failed at startup. No-op when the row exists. * fix(state): release lock between context queries in search_messages The context-window queries (one per FTS5 match) were running inside the same lock acquisition as the primary FTS5 query, holding the lock for O(N) sequential SQLite round-trips. Move per-match context fetches outside the outer lock block so each acquires the lock independently, keeping critical sections short and allowing other threads to interleave. * fix(session): prefer longer source in load_transcript to prevent legacy truncation When a long-lived session pre-dates SQLite storage (e.g. sessions created before the DB layer was introduced, or after a clean deployment that reset the DB), _flush_messages_to_session_db only writes the *new* messages from the current turn to SQLite — it skips messages already present in conversation_history, assuming they are already persisted. That assumption fails for legacy JSONL-only sessions: Turn N (first after DB migration): load_transcript(id) → SQLite: 0 → falls back to JSONL: 994 ✓ _flush_messages_to_session_db: skip first 994, write 2 new → SQLite: 2 Turn N+1: load_transcript(id) → SQLite: 2 → returns immediately ✗ Agent sees 2 messages of history instead of 996 The same pattern causes the reported symptom: session JSON truncated to 4 messages (_save_session_log writes agent.messages which only has 2 history + 2 new = 4). Fix: always load both sources and return whichever is longer. For a fully-migrated session SQLite will always be ≥ JSONL, so there is no regression. For a legacy session that hasn't been bootstrapped yet, JSONL wins and the full history is restored. Closes #3212 * test: add load_transcript source preference tests for #3212 Covers: JSONL longer returns JSONL, SQLite longer returns SQLite, SQLite empty falls back to JSONL, both empty returns empty, equal length prefers SQLite (richer reasoning fields). --------- Co-authored-by: Mibayy <mibayy@hermes.ai> Co-authored-by: kewe63 <kewe.3217@gmail.com> Co-authored-by: Mibayy <mibayy@users.noreply.github.com>
2026-03-26 13:47:14 -07:00
db_messages = []
# Try SQLite first
if self._db:
try:
fix(state): SQLite concurrency hardening + session transcript integrity (#3249) * fix(session-db): survive CLI/gateway concurrent write contention Closes #3139 Three layered fixes for the scenario where CLI and gateway write to state.db concurrently, causing create_session() to fail with 'database is locked' and permanently disabling session_search on the gateway side. 1. Increase SQLite connection timeout: 10s -> 30s hermes_state.py: longer window for the WAL writer to finish a batch flush before the other process gives up entirely. 2. INSERT OR IGNORE in create_session hermes_state.py: prevents IntegrityError on duplicate session IDs (e.g. gateway restarts while CLI session is still alive). 3. Don't null out _session_db on create_session failure (main fix) run_agent.py: a transient lock at agent startup must not permanently disable session_search for the lifetime of that agent instance. _session_db now stays alive so subsequent flushes and searches work once the lock clears. 4. New ensure_session() helper + call it during flush hermes_state.py: INSERT OR IGNORE for a minimal session row. run_agent.py _flush_messages_to_session_db: calls ensure_session() before appending messages, so the FK constraint is satisfied even when create_session() failed at startup. No-op when the row exists. * fix(state): release lock between context queries in search_messages The context-window queries (one per FTS5 match) were running inside the same lock acquisition as the primary FTS5 query, holding the lock for O(N) sequential SQLite round-trips. Move per-match context fetches outside the outer lock block so each acquires the lock independently, keeping critical sections short and allowing other threads to interleave. * fix(session): prefer longer source in load_transcript to prevent legacy truncation When a long-lived session pre-dates SQLite storage (e.g. sessions created before the DB layer was introduced, or after a clean deployment that reset the DB), _flush_messages_to_session_db only writes the *new* messages from the current turn to SQLite — it skips messages already present in conversation_history, assuming they are already persisted. That assumption fails for legacy JSONL-only sessions: Turn N (first after DB migration): load_transcript(id) → SQLite: 0 → falls back to JSONL: 994 ✓ _flush_messages_to_session_db: skip first 994, write 2 new → SQLite: 2 Turn N+1: load_transcript(id) → SQLite: 2 → returns immediately ✗ Agent sees 2 messages of history instead of 996 The same pattern causes the reported symptom: session JSON truncated to 4 messages (_save_session_log writes agent.messages which only has 2 history + 2 new = 4). Fix: always load both sources and return whichever is longer. For a fully-migrated session SQLite will always be ≥ JSONL, so there is no regression. For a legacy session that hasn't been bootstrapped yet, JSONL wins and the full history is restored. Closes #3212 * test: add load_transcript source preference tests for #3212 Covers: JSONL longer returns JSONL, SQLite longer returns SQLite, SQLite empty falls back to JSONL, both empty returns empty, equal length prefers SQLite (richer reasoning fields). --------- Co-authored-by: Mibayy <mibayy@hermes.ai> Co-authored-by: kewe63 <kewe.3217@gmail.com> Co-authored-by: Mibayy <mibayy@users.noreply.github.com>
2026-03-26 13:47:14 -07:00
db_messages = self._db.get_messages_as_conversation(session_id)
except Exception as e:
logger.debug("Could not load messages from DB: %s", e)
fix(state): SQLite concurrency hardening + session transcript integrity (#3249) * fix(session-db): survive CLI/gateway concurrent write contention Closes #3139 Three layered fixes for the scenario where CLI and gateway write to state.db concurrently, causing create_session() to fail with 'database is locked' and permanently disabling session_search on the gateway side. 1. Increase SQLite connection timeout: 10s -> 30s hermes_state.py: longer window for the WAL writer to finish a batch flush before the other process gives up entirely. 2. INSERT OR IGNORE in create_session hermes_state.py: prevents IntegrityError on duplicate session IDs (e.g. gateway restarts while CLI session is still alive). 3. Don't null out _session_db on create_session failure (main fix) run_agent.py: a transient lock at agent startup must not permanently disable session_search for the lifetime of that agent instance. _session_db now stays alive so subsequent flushes and searches work once the lock clears. 4. New ensure_session() helper + call it during flush hermes_state.py: INSERT OR IGNORE for a minimal session row. run_agent.py _flush_messages_to_session_db: calls ensure_session() before appending messages, so the FK constraint is satisfied even when create_session() failed at startup. No-op when the row exists. * fix(state): release lock between context queries in search_messages The context-window queries (one per FTS5 match) were running inside the same lock acquisition as the primary FTS5 query, holding the lock for O(N) sequential SQLite round-trips. Move per-match context fetches outside the outer lock block so each acquires the lock independently, keeping critical sections short and allowing other threads to interleave. * fix(session): prefer longer source in load_transcript to prevent legacy truncation When a long-lived session pre-dates SQLite storage (e.g. sessions created before the DB layer was introduced, or after a clean deployment that reset the DB), _flush_messages_to_session_db only writes the *new* messages from the current turn to SQLite — it skips messages already present in conversation_history, assuming they are already persisted. That assumption fails for legacy JSONL-only sessions: Turn N (first after DB migration): load_transcript(id) → SQLite: 0 → falls back to JSONL: 994 ✓ _flush_messages_to_session_db: skip first 994, write 2 new → SQLite: 2 Turn N+1: load_transcript(id) → SQLite: 2 → returns immediately ✗ Agent sees 2 messages of history instead of 996 The same pattern causes the reported symptom: session JSON truncated to 4 messages (_save_session_log writes agent.messages which only has 2 history + 2 new = 4). Fix: always load both sources and return whichever is longer. For a fully-migrated session SQLite will always be ≥ JSONL, so there is no regression. For a legacy session that hasn't been bootstrapped yet, JSONL wins and the full history is restored. Closes #3212 * test: add load_transcript source preference tests for #3212 Covers: JSONL longer returns JSONL, SQLite longer returns SQLite, SQLite empty falls back to JSONL, both empty returns empty, equal length prefers SQLite (richer reasoning fields). --------- Co-authored-by: Mibayy <mibayy@hermes.ai> Co-authored-by: kewe63 <kewe.3217@gmail.com> Co-authored-by: Mibayy <mibayy@users.noreply.github.com>
2026-03-26 13:47:14 -07:00
# Load legacy JSONL transcript (may contain more history than SQLite
# for sessions created before the DB layer was introduced).
transcript_path = self.get_transcript_path(session_id)
fix(state): SQLite concurrency hardening + session transcript integrity (#3249) * fix(session-db): survive CLI/gateway concurrent write contention Closes #3139 Three layered fixes for the scenario where CLI and gateway write to state.db concurrently, causing create_session() to fail with 'database is locked' and permanently disabling session_search on the gateway side. 1. Increase SQLite connection timeout: 10s -> 30s hermes_state.py: longer window for the WAL writer to finish a batch flush before the other process gives up entirely. 2. INSERT OR IGNORE in create_session hermes_state.py: prevents IntegrityError on duplicate session IDs (e.g. gateway restarts while CLI session is still alive). 3. Don't null out _session_db on create_session failure (main fix) run_agent.py: a transient lock at agent startup must not permanently disable session_search for the lifetime of that agent instance. _session_db now stays alive so subsequent flushes and searches work once the lock clears. 4. New ensure_session() helper + call it during flush hermes_state.py: INSERT OR IGNORE for a minimal session row. run_agent.py _flush_messages_to_session_db: calls ensure_session() before appending messages, so the FK constraint is satisfied even when create_session() failed at startup. No-op when the row exists. * fix(state): release lock between context queries in search_messages The context-window queries (one per FTS5 match) were running inside the same lock acquisition as the primary FTS5 query, holding the lock for O(N) sequential SQLite round-trips. Move per-match context fetches outside the outer lock block so each acquires the lock independently, keeping critical sections short and allowing other threads to interleave. * fix(session): prefer longer source in load_transcript to prevent legacy truncation When a long-lived session pre-dates SQLite storage (e.g. sessions created before the DB layer was introduced, or after a clean deployment that reset the DB), _flush_messages_to_session_db only writes the *new* messages from the current turn to SQLite — it skips messages already present in conversation_history, assuming they are already persisted. That assumption fails for legacy JSONL-only sessions: Turn N (first after DB migration): load_transcript(id) → SQLite: 0 → falls back to JSONL: 994 ✓ _flush_messages_to_session_db: skip first 994, write 2 new → SQLite: 2 Turn N+1: load_transcript(id) → SQLite: 2 → returns immediately ✗ Agent sees 2 messages of history instead of 996 The same pattern causes the reported symptom: session JSON truncated to 4 messages (_save_session_log writes agent.messages which only has 2 history + 2 new = 4). Fix: always load both sources and return whichever is longer. For a fully-migrated session SQLite will always be ≥ JSONL, so there is no regression. For a legacy session that hasn't been bootstrapped yet, JSONL wins and the full history is restored. Closes #3212 * test: add load_transcript source preference tests for #3212 Covers: JSONL longer returns JSONL, SQLite longer returns SQLite, SQLite empty falls back to JSONL, both empty returns empty, equal length prefers SQLite (richer reasoning fields). --------- Co-authored-by: Mibayy <mibayy@hermes.ai> Co-authored-by: kewe63 <kewe.3217@gmail.com> Co-authored-by: Mibayy <mibayy@users.noreply.github.com>
2026-03-26 13:47:14 -07:00
jsonl_messages = []
if transcript_path.exists():
with open(transcript_path, "r", encoding="utf-8") as f:
for line in f:
line = line.strip()
if line:
try:
jsonl_messages.append(json.loads(line))
except json.JSONDecodeError:
logger.warning(
"Skipping corrupt line in transcript %s: %s",
session_id, line[:120],
)
# Prefer whichever source has more messages.
#
# Background: when a session pre-dates SQLite storage (or when the DB
# layer was added while a long-lived session was already active), the
# first post-migration turn writes only the *new* messages to SQLite
# (because _flush_messages_to_session_db skips messages already in
# conversation_history, assuming they're persisted). On the *next*
# turn load_transcript returns those few SQLite rows and ignores the
# full JSONL history — the model sees a context of 1-4 messages instead
# of hundreds. Using the longer source prevents this silent truncation.
if len(jsonl_messages) > len(db_messages):
if db_messages:
logger.debug(
"Session %s: JSONL has %d messages vs SQLite %d"
"using JSONL (legacy session not yet fully migrated)",
session_id, len(jsonl_messages), len(db_messages),
)
return jsonl_messages
return db_messages
def build_session_context(
source: SessionSource,
config: GatewayConfig,
session_entry: Optional[SessionEntry] = None
) -> SessionContext:
"""
Build a full session context from a source and config.
This is used to inject context into the agent's system prompt.
"""
connected = config.get_connected_platforms()
home_channels = {}
for platform in connected:
home = config.get_home_channel(platform)
if home:
home_channels[platform] = home
context = SessionContext(
source=source,
connected_platforms=connected,
home_channels=home_channels,
fix(gateway): preserve sender attribution in shared group sessions Generalize shared multi-user session handling so non-thread group sessions (group_sessions_per_user=False) get the same treatment as shared threads: inbound messages are prefixed with [sender name], and the session prompt shows a multi-user note instead of pinning a single **User:** line into the cached system prompt. Before: build_session_key already treated these as shared sessions, but _prepare_inbound_message_text and build_session_context_prompt only recognized shared threads — creating cross-user attribution drift and prompt-cache contamination in shared groups. - Add is_shared_multi_user_session() helper alongside build_session_key() so both the session key and the multi-user branches are driven by the same rules (DMs never shared, threads shared unless thread_sessions_per_user, groups shared unless group_sessions_per_user). - Add shared_multi_user_session field to SessionContext, populated by build_session_context() from config. - Use context.shared_multi_user_session in the prompt builder (label is 'Multi-user thread' when a thread is present, 'Multi-user session' otherwise). - Use the helper in _prepare_inbound_message_text so non-thread shared groups also get [sender] prefixes. Default behavior unchanged: DMs stay single-user, groups with group_sessions_per_user=True still show the user normally, shared threads keep their existing multi-user behavior. Tests (65 passed): - tests/gateway/test_session.py: new shared non-thread group prompt case. - tests/gateway/test_shared_group_sender_prefix.py: inbound preprocessing for shared non-thread groups and default groups.
2026-04-21 00:37:26 -07:00
shared_multi_user_session=is_shared_multi_user_session(
source,
group_sessions_per_user=getattr(config, "group_sessions_per_user", True),
thread_sessions_per_user=getattr(config, "thread_sessions_per_user", False),
),
)
if session_entry:
context.session_key = session_entry.session_key
context.session_id = session_entry.session_id
context.created_at = session_entry.created_at
context.updated_at = session_entry.updated_at
return context