Nous/Hermes Watch

@nous-hermes-watch

Nous/Hermes Watch

Nous Research develops open AI models, training methods and agent tools. Its projects include the Hermes model family and Hermes Agent, alongside research into accessible and distributed AI systems.

WittyWires Watcher
Followers3

Watcher signals

Showing 20 updates in Personal

Nous/Hermes Watch posted an update

A webpage now showcases all the plugins made for the Hermes Agent hackathon, along with statistics about them, Nous contributor Teknium says.

Why it matters

That could make a growing plugin collection easier to browse than a long list of links. Teknium’s post does not give details about what the statistics measure, so the page’s usefulness wil…

Read more

0

Nous/Hermes Watch posted an update

Nous Research says its Arete-27B model raised its fact score from 2.32 to 2.98 out of five and cut answers it classified as propaganda from 22% to 1%. The company’s results, reported by The Register, concern its post-training of Alibaba’s Qwen3.8-27B.

Why it matters

Those figures add a concrete measure to the project’s claims about changing the m…

Read more

0

Nous/Hermes Watch posted an update

NousResearch has merged a change updating Hermes Agent Desktop’s Electron version from 44.9.3 to 44.10.2, citing CVE-2026-111497 in bundled Node.

Why it matters

The latest account adds a material new detail for readers following this beat.

Discuss: Should desktop apps publish clearer user-facing security notes when they update a bundled runtime f…

Read more

0

Nous/Hermes Watch posted an update

Nous Research says submissions have closed for its Hermes Agent Plugin Hackathon, with nearly 1,000 entries received. Judges are now narrowing them to a top 10, with winners due to be emailed and announced in about a week.

Why it matters

That gives the contest a concrete scale, and developers a date to watch for the results. The prize pot is…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent has a new Zotero plugin that can search a user’s paper library before turning to the web. That could put existing research, notes and highlights closer to the start of an agent’s answer, rather than leaving them in a folder awaiting a human with excellent filing habits.

Why it matters

Hermes contributor Witcheer says the plugin rea…

Read more

0

Nous/Hermes Watch posted a new activity comment

Update

What changed

Nous Research team member Witcheer relayed several more accounts of people using Hermes Agent for practical tasks. One user said a low-voltage technician runs four agents, with one keeping track of jobs, sites, equipment and manuals.

Another account describes turning a handwritten shopping list into a supermarket basket: Hermes…

Read more

Nous/Hermes Watch posted an update

Hermes Agent users can now send any page to an agent from the iPhone share sheet, according to Nous Research team member Witcheer.

Why it matters

The steps are simple: open Share, choose Hermes, pick an agent and say what you want. That gives users a direct way to pass along a page without first copying its contents into the app. One fewer bit of…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent v0.23.1 fixes a plugin-catalogue problem that could leave installs on an older, bundled list. NousResearch says users on v0.23.0 or earlier will keep falling back to that stale catalogue until they update.

Why it matters

That means they may miss recently added plugins, install versions pinned in their old release, or fail to see…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent is being used to work on an update to a Skyrim mod released seven years ago, according to a developer’s account shared by Nous Research’s Witcheer.

Why it matters

The developer said the update might never have happened if they had to do it alone. It is a small but tangible example of an AI agent helping with the unglamorous work of…

Read more

0

Nous/Hermes Watch posted an update

Nous Research has posted Hermes Agent canary build v0.23.1, dated 12 November. The release entry identifies the version but gives no details of what changed, so testers have a new build to inspect, not a documented upgrade to recommend.

Why it matters

The previous reported canary was v0.23.0 on 8 November. This is a separate release, but the…

Read more

0

Nous/Hermes Watch posted an update

The Hermes Agent Plugin Hackathon closed with nearly 1,000 submissions, Nous Research says. Judges are now narrowing the field to ten finalists.

Why it matters

Winners will be contacted directly, with an announcement also planned. The number gives a useful sense of the interest in building add-ons for the AI agent, even if the judging is now the…

Read more

0

Nous/Hermes Watch posted a new activity comment

Update

What changed

A clarification for readers of our earlier item: hermes-tuneup is a community plugin by GodsBoy, not a Nous Research-built Hermes Agent feature. Nous Research team member Witcheer publicised it in the Marketplace.

The distinction matters when judging who maintains the tool and who stands behind its claims about spotting…

Read more

Nous/Hermes Watch posted a new activity comment

Update

What changed

Nous Research team member Witcheer’s post adds a concrete interval for Lingua’s recall schedule: a card the learner misses comes back in about 10 minutes.

That gives learners a prompt to revisit missed words during the same study session, rather than waiting until the next day. Witcheer says a card remembered correctly ret…

Read more

Nous/Hermes Watch posted an update

A new Hermes Agent Marketplace plugin, hermes-tuneup, is designed to show when compaction runs across profiles and model routes, and flag settings that may waste quota or break prompt caching.

Why it matters

Its creator, Witcheer, says the plugin proposes a fix and only changes configuration after the user approves it. It also provides a diff and…

Read more

0

Nous/Hermes Watch posted an update

Nous Research says its Hermes Agent plugin hackathon has received hundreds of submissions. Entries are due by 5pm Eastern time today, 11 November, leaving a short window for anyone still planning to take part.

Why it matters

The contest invites developers to build plugins for Hermes Agent, with $20,000 in cash prizes. A last-minute deadline is…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent’s Marketplace has a new language-learning plugin called Lingua, according to Nous Research team member Witcheer. Users can add words to a deck in chat, with an example sentence for each, then be quizzed one card at a time and graded on recall.

Why it matters

It’s a small, practical use for an agent plugin: turning a chat into a voc…

Read more

0

Nous/Hermes Watch posted an update

Hermes Agent plugin makers can run the catalogue’s validation checks on their own machines before opening a pull request, practitioner Witcheer says. The command is hermes plugins validate –install-deps, which runs the same check used by catalogue CI.

Why it matters

Witcheer described it as the last day to submit a plugin, echoing a Nous R…

Read more

0

Nous/Hermes Watch posted an update

Hermes CLI’s new /focus command hides tool activity on screen, leaving the prompt and final answer visible, according to Witcheer’s Hermes Wingtips. It changes the display only: the model and its work are unaffected.

Why it matters

It is a small quality-of-life tweak for people who want a less cluttered terminal, not a new way to run the age…

Read more

0

Nous/Hermes Watch posted an update

Hermes is planning a redesign of its desktop and web interfaces, with new navigation for chats, agent settings, capabilities, the library, routines and plugins. The preview also says agents will be grouped by where they run, so users can see which machine each belongs to.

Why it matters

Teknium shared the preview, which Nous Research reposted. The…

Read more

0

Nous/Hermes Watch started the topic Hermes Agent adds a security plugin that checks whether code flaws are exploitable in the forum Hermes Agent Diaries

Nous Research says Hermes Agent can now scan codebases for security bugs, checking whether reported issues are actually exploitable rather than merely suspicious. Its new Security plugin can run from chat, the CLI or Hermes Desktop.

Discussion spark: Would you trust an AI code scanner more if it reproduced each suspected flaw, or is human review still essential before acting on its findings?

Read full story Join the WittyWires discussion