MEMIND_SESSION_REPLY_IDLE_TIMEOUT_MS (off by default, min 30s) fails a reply
that stops streaming while no tool is running, and is recoverable like other
finish errors. The goose Finish reason is surfaced in session_finished.
Co-authored-by: Cursor <cursoragent@cursor.com>
Close the chat-session dead zone without flipping production flags: record 0.72 fallbacks, stop C-2 from treating「先聊聊」as execute, and promote「不对你去执行」after a completed direct turn on both h5direct and date sessions.
Co-authored-by: Cursor <cursoragent@cursor.com>
Wire WECHAT_AGENT_REPLY_TIMEOUT into the same Cursor takeover path used by
agent runs, raise the default WeChat reply timeout to 20 minutes, and fall
back to env-based tool-gateway detection when runtime exports are missing.
Co-authored-by: Cursor <cursoragent@cursor.com>
When a page-generation run times out without recoverable HTML, retry once
via Cursor code executor if MEMIND_CURSOR_PAGE_TIMEOUT_TAKEOVER is enabled.
Also release delivery contracts after successful deliverable recovery.
Co-authored-by: Cursor <cursoragent@cursor.com>
Wire DeepSeek Harness as an optional fourth code executor (default off)
and add a loopback observation script that compares prompt tokens through
headroom proxy with OPENAI_TARGET_API_URL pointed at the compat proxy.
Co-authored-by: Cursor <cursoragent@cursor.com>
Introduce RRF-based recall fusion for personal/episodic/temporal paths,
zvec-grep workspace shadow probing for code executors, and a local
headroom proxy starter aligned with the fusion plan defaults.
Co-authored-by: Cursor <cursoragent@cursor.com>
Extend h5_experience with structured fields, wire mindspace-agent-runner and agent-run-gateway to persist task_outcome records with provenance, and add local migration and verification scripts.
Co-authored-by: Cursor <cursoragent@cursor.com>
Phase B appends snapshot context when leaving h5direct sessions so Goose no longer starts blind. Phase C lets the 0.72 fallback return null on gated chat text so the existing LLM router can run, flags default off.
Co-authored-by: Cursor <cursoragent@cursor.com>
Keep correction follow-ups on Agent using session transcript and active task
context, and let Page Data validation reach Orchestrator when mode is active
instead of requiring shadowEngine.
Co-authored-by: Cursor <cursoragent@cursor.com>
Parse Cursor CLI usage from stream-json output and charge through billSessionUsage so subscription quota is consumed first and wallet overage applies when needed.
Co-authored-by: Cursor <cursoragent@cursor.com>
Recover requiredExecutor fallback, direct cursor launch in tool gateway,
and user-facing brand sanitization in source so the next portal release
can replace manual bundled edits on 103.
Co-authored-by: Cursor <cursoragent@cursor.com>
Persist goal runs in MySQL, bind agent runs to checkpoints, expose awaiting-approval
UX in chat, and add admin inspection routes with local verify scripts.
Co-authored-by: Cursor <cursoragent@cursor.com>
Map Page Data failure codes to Chinese remediation hints, trigger one-shot
goosed repair for remediable cases, and recover poisoned thinking sessions.
Route local DeepSeek through the no-think proxy via host.docker.internal so
tool rounds no longer hit reasoning_content 400; document gate and case-study
scenarios for event registration repair.
Co-authored-by: Cursor <cursoragent@cursor.com>
Hook shadow pipeline observation into session finish and agent runs, and
expose skill runtime settings through auth and runtime status endpoints.
Co-authored-by: Cursor <cursoragent@cursor.com>
john