seb/hermes-agent
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(slack): make video attachments available to agents (#45512)

Browse Source
This commit is contained in:
Teknium
2026-06-13 03:33:27 -07:00
committed by GitHub
parent 197337cc47
commit 2a5dc0ef3d
6 changed files with 274 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gateway/platforms/slack.py
+113 -1
View File
@@ -45,10 +45,12 @@ from gateway.platforms.base import (
ProcessingOutcome,
SendResult,
SUPPORTED_DOCUMENT_TYPES,
SUPPORTED_VIDEO_TYPES,
is_host_excluded_by_no_proxy,
resolve_proxy_url,
safe_url_for_log,
cache_document_from_bytes,
cache_video_from_bytes,
)
@@ -880,7 +882,7 @@ class SlackAdapter(BasePlatformAdapter):
# doesn't log noisy 404 "unhandled request" warnings.
@self._app.event("file_shared")
async def handle_file_shared(event, say):
pass
await self._handle_slack_file_shared(event)
@self._app.event("file_created")
async def handle_file_created(event, say):
@@ -2173,6 +2175,84 @@ class SlackAdapter(BasePlatformAdapter):
self._cache_assistant_thread_metadata(metadata)
self._seed_assistant_thread_session(metadata)
async def _handle_slack_file_shared(self, event: dict) -> None:
"""Fallback for Slack file shares that do not arrive as message.files.
Slack documents ``file_shared`` as a file-ID-only event; callers must
fetch ``files.info`` to get the file object. Keep this intentionally
narrow: normal image/audio/document uploads already arrive on the
message event, but some video shares have only been observed through
this lifecycle event.
"""
channel_id = event.get("channel_id") or event.get("channel") or ""
file_id = event.get("file_id") or (event.get("file") or {}).get("id") or ""
if not channel_id or not file_id:
return
team_id = event.get("team_id") or event.get("team") or ""
try:
client = self._team_clients.get(team_id) if team_id else None
info_resp = await (client or self._get_client(channel_id)).files_info(
file=file_id
)
except Exception as exc:
response = getattr(exc, "response", None)
detail = self._describe_slack_api_error(response, file_obj={"id": file_id})
logger.warning("[Slack] files.info error for file_shared %s: %s", file_id, detail or exc)
return
if not info_resp.get("ok"):
detail = self._describe_slack_api_error(info_resp, file_obj={"id": file_id})
logger.warning(
"[Slack] files.info failed for file_shared %s: %s",
file_id,
detail or info_resp.get("error"),
)
return
file_obj = info_resp.get("file") or {}
if not str(file_obj.get("mimetype", "")).startswith("video/"):
return
share = None
for bucket in (file_obj.get("shares") or {}).values():
if not isinstance(bucket, dict):
continue
channel_shares = bucket.get(channel_id)
if channel_shares:
share = channel_shares[0]
break
if share is None:
for shares in bucket.values():
if shares:
share = shares[0]
break
share = share or {}
ts = share.get("ts") or event.get("event_ts") or ""
thread_ts = share.get("thread_ts") or ""
# Give Slack's normal message.file_share event a chance to arrive first.
# If it does, _handle_slack_message records the same share ts and this
# fallback skips instead of duplicating the user turn.
await asyncio.sleep(0.75)
if ts and self._dedup.is_duplicate(ts):
return
fallback_event = {
"type": "message",
"subtype": "file_share",
"text": "",
"user": event.get("user_id") or file_obj.get("user", ""),
"channel": channel_id,
"channel_type": "im" if channel_id.startswith("D") else "channel",
"team": team_id,
"ts": "", # already recorded above; avoid tripping our own dedup guard
"files": [file_obj],
}
if thread_ts and thread_ts != ts:
fallback_event["thread_ts"] = thread_ts
await self._handle_slack_message(fallback_event)
async def _handle_slack_message(self, event: dict) -> None:
"""Handle an incoming Slack message event."""
# Dedup: Slack Socket Mode can redeliver events after reconnects (#4777)
@@ -2572,6 +2652,36 @@ class SlackAdapter(BasePlatformAdapter):
e,
exc_info=True,
)
elif mimetype.startswith("video/") and url:
try:
original_filename = f.get("name", "")
_, ext = os.path.splitext(original_filename)
ext = ext.lower()
if ext not in SUPPORTED_VIDEO_TYPES:
mime_to_ext = {v: k for k, v in SUPPORTED_VIDEO_TYPES.items()}
ext = mime_to_ext.get(
mimetype.split(";", 1)[0].lower(), ".mp4"
)
raw_bytes = await self._download_slack_file_bytes(
url, team_id=team_id
)
cached_path = cache_video_from_bytes(raw_bytes, ext=ext)
media_urls.append(cached_path)
media_types.append(SUPPORTED_VIDEO_TYPES.get(ext, mimetype))
logger.debug("[Slack] Cached user video: %s", cached_path)
except Exception as e: # pragma: no cover - defensive logging
detail = self._describe_slack_download_failure(e, file_obj=f)
if detail:
attachment_notices.append(detail)
logger.warning("[Slack] %s", detail)
else:
logger.warning(
"[Slack] Failed to cache video from %s: %s",
url,
e,
exc_info=True,
)
elif url:
# Try to handle as a document attachment
try:
@@ -2666,6 +2776,8 @@ class SlackAdapter(BasePlatformAdapter):
if msg_type != MessageType.COMMAND and media_types:
if any(m.startswith("image/") for m in media_types):
msg_type = MessageType.PHOTO
elif any(m.startswith("video/") for m in media_types):
msg_type = MessageType.VIDEO
elif any(m.startswith("audio/") for m in media_types):
msg_type = MessageType.VOICE
else:
gateway/run.py
+24
View File
@@ -1420,6 +1420,8 @@ def _build_media_placeholder(event) -> str:
parts.append(f"[User sent an image: {url}]")
elif mtype.startswith("audio/"):
parts.append(f"[User sent audio: {url}]")
elif mtype.startswith("video/") or getattr(event, "message_type", None) == MessageType.VIDEO:
parts.append(f"[User sent a video: {url}]")
else:
parts.append(f"[User sent a file: {url}]")
return "\n".join(parts)
@@ -7637,6 +7639,7 @@ class GatewayRunner(GatewayAuthorizationMixin, GatewayKanbanWatchersMixin, Gatew
# Declare at outer scope so the audio-file-paths handling block below
# remains safe when ``event.media_urls`` is empty (no inner block runs).
audio_file_paths: list[str] = []
video_paths: list[str] = []
if event.media_urls:
image_paths = []
@@ -7654,6 +7657,8 @@ class GatewayRunner(GatewayAuthorizationMixin, GatewayKanbanWatchersMixin, Gatew
and event.message_type not in {MessageType.AUDIO, MessageType.DOCUMENT}
):
audio_paths.append(path)
if mtype.startswith("video/") or event.message_type == MessageType.VIDEO:
video_paths.append(path)
if image_paths:
# Decide routing: native (attach pixels) vs text (vision_analyze
@@ -7752,6 +7757,25 @@ class GatewayRunner(GatewayAuthorizationMixin, GatewayKanbanWatchersMixin, Gatew
)
message_text = f"{_note}\n\n{message_text}"
if video_paths:
from tools.credential_files import to_agent_visible_cache_path as _to_agent_path
for _vpath in video_paths:
_basename = os.path.basename(_vpath)
_parts = _basename.split("_", 2)
_display = _parts[2] if len(_parts) >= 3 else _basename
_display = re.sub(r'[^\w.\- ]', '_', _display)
_agent_path = _to_agent_path(_vpath)
_note = (
f"[The user sent a video attachment: '{_display}'. "
f"It is saved at: {_agent_path}. "
f"Its content is not inlined here. If the user's request involves "
f"what the video contains, inspect or process it yourself — for "
f"example by passing the path to a video analysis or media tool — "
f"instead of asking the user to describe it. Only ask what to do "
f"with it if their intent is genuinely unclear.]"
)
message_text = f"{_note}\n\n{message_text}"
if event.media_urls and event.message_type == MessageType.DOCUMENT:
import mimetypes as _mimetypes
from tools.credential_files import to_agent_visible_cache_path