Grok Build just got another big update, taking another major step toward becoming a full agent workspace
Desktop app tools now connect through a dedicated first-party MCP server, a new long-running workspace daemon can expose folders to Computer Hub, finished subagents can be continued instead of restarted, and background agents now return their actual output directly
Add better diff previews, smarter tool priority, stronger safety around destructive Git commands, and faster session startup
Release Notes: v1.0.22
Features: • Built-in agent tools now take precedence over user MCP servers when tool names collide. • Desktop app tools are now available through a dedicated first-party MCP server. • Remote control pickers can now group servers by device using announced device identity. • edit_file diffs now show real file line numbers instead of starting at line 1. • Permission prompts for file edits now auto-expand the diff row so you can see the change before approving. • Sending a message to a finished subagent now continues that same subagent instead of failing. • Background subagent completion messages now show the actual output instead of just a pointer to run the tool. • Dashboard now shows a unified header with location picker and an actions row for creating agents or resuming previous sessions. • New `grok-workspaced daemon` mode lets a long-lived process expose folders to the Computer Hub (separate from the Desktop-supervised sidecar).
Bug Fixes: • MCP connectors now correctly show when they need re-authentication even if no tools are listed. • Extensions modal shortcuts now work when the tab bar is focused, and invalid actions on headers show helpful messages. • Send now button and shortcuts now work correctly when an automatic background turn is running. • Resumed sessions now show the exact text you typed for mid-turn follow-ups instead of the wrapped system message. • `/skills` command now updates the model with newly installed skills even when written from another NFS client. • Auto mode no longer instantly runs destructive `git checkout --` commands; they now go through the model for safety. • Fixed a crash that could occur when restoring pasted content after certain skill injections or rewinds. • Subagents and workflows now respect the same bash timeout and auto-background settings configured for the parent session. • Fixed model backend selection so an explicit chat_completions setting in config.toml is no longer overwritten by same-slug siblings. • Slash command advertisements no longer spam repeatedly when your current directory is your home folder. • Foreground shell commands no longer produce spurious background-task reminders or incorrect kill results. • Queued follow-up buttons now appear in the order [Send now][edit][cancel]. • `/compact instructions` in the pager are now passed to the compaction backend (bare `/compact` is unchanged). • `/goal` (and resume) after pressing Esc now correctly starts the goal planner instead of failing with "Planning failed". • run_terminal_cmd now correctly tells the model that foreground commands are backgrounded after ~15s instead of the 120s timeout value. • Skill announcements no longer repeat the same list of available skills multiple times in a session. • Monitor events in multi-session processes now correctly wake idle sessions instead of being buffered until the next prompt.
Performance: • Opening or resuming sessions no longer pauses the interface while reading MCP config files.
5·CNeutral
n
news9/6news
Bittensor Ecosystem Highlights :: August 31–September 6, 2026
This week’s biggest stories across Bittensor came from Gittensor, Score, Good Morning, OpenRoboto, Bitsec and Trishool.
[ @gittensor_io - Subnet 74 ] Gittensor released a Qwen3.8-27B checkpoint built to run on a single RTX 5090, and it has now passed 500K downloads on Hugging Face.
The model runs on a single consumer GPU, making a powerful open model much easier for developers to run on their own hardware.
[ @webuildscore - Subnet 44 ] Score expanded its fuel-station rollout beyond Avia, adding Shell and Eni locations. The first deployments are with franchisees, with the potential to expand into direct contracts with Shell and Eni after the initial rollout.
[ @say_gm_ - Subnet 28 ] Good Morning made OpenAI’s GPT-6 Astra available through Bittensor, with private, verifiable access at an 8.3% discount. Gm had Astra before OpenRouter, a strong distribution win for Subnet 28.
[ @openroboto - Subnet 80 ] OpenRoboto made @robbyant_brain’s LingBot-VLA 2.0 its new base model for community post-training on Subnet 80. Robbyant welcomed the move directly, praising OpenRoboto for pushing the model further.
OpenRoboto also brought its xArm 6 setup online, preparing to validate winning simulation models on a real robot.
[ @bitsecai - Subnet 60 ] Bitsec put Sentios, its continuous code-auditing product, into one public site with pricing and coverage details.
Instead of a one-off security audit, Sentios keeps checking code as it changes, using Subnet 60’s security agents to keep searching for new vulnerabilities.
[ @trishoolai - Subnet 23 ] Trishool was accepted into OpenAI’s Trusted Access for Cyber program, which gives vetted security teams access to advanced frontier capabilities for defensive work. The team says this will help bring Halo, its AI security model, to more organizations as it expands commercially.
75·ALong
n
news9/6news
Bittensor Ecosystem Highlights :: August 31–September 6, 2026
This week’s biggest stories across Bittensor came from Gittensor, Score, Good Morning, OpenRoboto, Bitsec and Trishool.
[ @gittensor_io - Subnet 74 ]
Gittensor released a Qwen3.8-27B checkpoint built to run on a single RTX 5090, and it has now passed 500K downloads on Hugging Face.
The model runs on a single consumer GPU, making a powerful open model much easier for developers to run on their own hardware.
[ @webuildscore - Subnet 44 ]
Score expanded its fuel-station rollout beyond Avia, adding Shell and Eni locations.
The first deployments are with franchisees, with the potential to expand into direct contracts with Shell and Eni after the initial rollout.
[ @say_gm_ - Subnet 28 ]
Good Morning made OpenAI’s GPT-6 Astra available through Bittensor, with private, verifiable access at an 8.3% discount.
Gm had Astra before OpenRouter, a strong distribution win for Subnet 28.
[ @openroboto - Subnet 80 ]
OpenRoboto made @robbyant_brain’s LingBot-VLA 2.0 its new base model for community post-training on Subnet 80.
Robbyant welcomed the move directly, praising OpenRoboto for pushing the model further.
OpenRoboto also brought its xArm 6 setup online, preparing to validate winning simulation models on a real robot.
[ @bitsecai - Subnet 60 ]
Bitsec put Sentios, its continuous code-auditing product, into one public site with pricing and coverage details.
Instead of a one-off security audit, Sentios keeps checking code as it changes, using Subnet 60’s security agents to keep searching for new vulnerabilities.
[ @trishoolai - Subnet 23 ]
Trishool was accepted into OpenAI’s Trusted Access for Cyber program, which gives vetted security teams access to advanced frontier capabilities for defensive work.
The team says this will help bring Halo, its AI security model, to more organizations as it expands commercially.
75·ALong
n
news9/4news
Why today's @USTreasury action against a Turkish bank matters more than the target's size suggests.
Golden Global Yatırım Bankası is small: licensed in 2019, two Istanbul branches, ~$500M in assets. But it grew 57% last year, multiplied its capital 11x, and sold $30M of dollar sukuk just four months ago. A tiny new bank growing that fast is either a great story or a laundromat. Here, Treasury says it was the latter.
It is also the first bank in a NATO ally hit under Operation Economic Outcast, Treasury's Iran campaign, and it comes a week after Egypt's Banque Misr UAE was cut off from U.S. banks. Secretary Bessent promised more every week. As Gulf channels close, Türkiye is the displacement route, and its banks will de-risk fast.
What did Golden Global do? Treasury says it was built to move Iran's China oil revenue into Türkiye for conversion to cash and gold, and to bank proxies of the IRGC's Qods Force in the network of Sitki Ayan, the Erdoğan-linked oil trader OFAC sanctioned in 2022.
Leaked records on @WikIran show the plumbing. Between November and December 2021, Baslam Nakliyat, Ayan's sanctioned front, sent €12M in six SWIFT payments into Golden Global shell-company accounts under "crude oil sale" assignment letters. Golden Global cleared the euros through Russia's Transkapitalbank, itself sanctioned months later for helping banks evade sanctions.